Dissertations / Theses on the topic 'Users’ preferences'

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2010.<br>Cataloged from PDF version of thesis.<br>Includes bibliographical references (p. 79).<br>Our goal is to construct a system that can determine a drivers preferences and goals and perform appropriate actions to aid the driver achieving his goals and improve the quality of his road behavior. Because the recommendation problem could be achieved effectively once we know the driver's intention, in this thesis, we are going to solve the problem to determine the driver's preferences. A supervised learning approach has already been applied to this problem. However, because the approach locally classify a small interval at a time and is memoryless, the supervised learning does not perform well on our goal. Instead, we need to introduce new approach which has following characteristics. First, it should consider the entire stream of measurements. Second, it should be tolerant to the environment. Third, it should be able to distinguish various intentions. In this thesis, two different approaches, Bayesian hypothesis testing and inverse reinforcement learning, will be used to classify and estimate the user's preferences. Bayesian hypothesis testing classifies the driver as one of several driving types. Assuming that the probability distributions of the features (i.e. average, standard deviation) for a short period of measurement are different among the driving types, Bayesian hypothesis testing classifies the driver as one of driving types by maintaining a belief distribution for each driving type and updating it online as more measurements are available. On the other hand, inverse reinforcement learning estimates the users' preferences as a linear combination of driving types. The inverse reinforcement learning approach assumes that the driver maximizes a reward function while driving, and his reward function is a linear combination of raw / expert features. Based on the observed trajectories of representative drivers, apprenticeship learning first calculates the reward function of each driving type with raw features, and these reward functions serve as expert features. After, with observed trajectories of a new driver, the same algorithm calculates the reward function of him, not with raw features, but with expert features, and estimates the preferences of any driver in a space of driving types.<br>by Jongu Shin.<br>M.Eng.

L'objectiu d'aquesta tesi és desenvolupar nous i diversos mètodes per modelar les preferències dels usuaris a les Xarxes Socials Online. Els mètodes proposats tenen com a finalitat ser aplicats en àrees de recerca com la Personalització o Recomanació d'ítems i la Detecció de Grups d'Usuaris amb gustos similars. Aquests mètodes poden ser agrupats en dos tipus: i) mètodes basats en tècniques d'anàlisi de textos (Part I, Capítols del 3 al 5) i ii) mètodes basats en teoria de grafs (Part II, Capítols 6 i 7). Amb els mètodes plantejats a la Part I és possible determinar el nivell d'interès dels usuaris en temes que són compartits en plataformes de microblogging. Hem pres com a cas d'estudi la participació digital de tweeters a la política. Els mètodes proposats a la Part II busquen definir un paper pels usuaris de les Xarxes Socials, ja sigui com a creadors o generadors de contingut i distribuïdors o consumidors de contingut. Hem plantejat un mètode on usuaris amb interessos similars però amb diferent rols són agrupats en una mateixa comunitat, de manera que els nous continguts es propaguen més ràpidament.<br>El objetivo de esta tesis es desarrollar nuevos y diversos métodos para modelar las preferencias de los usuarios en las Redes Sociales Online. Los métodos propuestos tienen como finalidad ser aplicados en áreas de investigación como la Personalización o Recomendación de ítems y la Detección de Grupos de Usuarios con gustos similares. Dichos métodos pueden ser agrupados en dos tipos: i) métodos basados en técnicas de análisis de texto (Parte I, Capítulos del 3 al 5) y ii) métodos basados en teoría de grafos (Parte II, Capítulos 6 y 7). Con los métodos planteados en la Parte I es posible determinar el nivel de interés de los usuarios en temas que son compartidos en plataformas de microblogging. Hemos tomado como caso de estudio la participación digital de tweeters en la política. Los métodos propuestos en la Parte II buscan definir un rol para los usuarios en Redes Sociales, ya sea como creadores o generadores de contenido y distribuidores o consumidores de contenido. Hemos planteado un método donde usuarios con intereses similares pero con distinto rol, son agrupados en una misma comunidad de forma que nuevo contenido se propague más rápidamente.<br>The objective of this thesis is to develop new and diverse methods to model the preferences of the users in the Online Social Networks. The proposed methods are intended to be applied in areas of research such as Personalization or Recommendation of items and the detection of groups of users with similar tastes. These methods can be grouped into two types: i) methods based on text analysis techniques (Chapters 3 to 5) and ii) methods based on graph theory (Chapters 6 and 7). With the methods proposed in i) it is possible to determine the level of interest of users on topics that are shared on microblogging platforms. We have taken as a case study the digital participation of tweeters in politics. The methods proposed in ii) seek to define a role for users in social networks, whether as creators or content generators and distributors or content consumers. We have proposed a method where users with similar interests but with different roles, are grouped in the same community so that new content spreads more quickly.

After compiling an appropriate list of beach criteria from established award programs and experts, the research landscape architect observed the Virginia Beach Resort Zone for areas of intense beach user activity.  The resulting analysis of these "hot-spots" indicated that urban resort beach users visiting the recreational beach during high-use times tended to gravitate toward locations on the resort beach which were close to three needs:  public parking, public restrooms, and inexpensive refreshments.  This pattern shows the apparent preferences of many beach users for certain amenities.<br>Master of Landscape Architecture

The introduction of an accounting standard requiring government departments to replace fund-type, cash-based accounting statements with business-type, accrual based accounting statements has led to criticism that business-type, general purpose financial statements do not take account of the information requirements of major users. Such criticism echoes a long standing debate in which the users of public sector financial statements and their informational requirements are analysed in competing models. One view suggests that there are many users with homogeneous informational needs, who can be classified into a few broad groups. The other view maintains that there are few users who have differential informational requirements. This research adds to the few empirical studies on the usefulness of public sector accounting statement information. The purpose of this research is to test the hypothesis that users perceive that there is no difference in the usefulness of fund-type, cash-based; business-type, accrual-based accounting statements, and both cash and accrual combined accounting statements. Responses from legislators, citizen/ interest group members, and preparers to a questionnaire provides the data for statistical analysis. Test results suggest that there is only moderate support for the hypothesis that heterogenous users have different information needs. Strong support is found for the hypothesis that combined sets of statements as opposed to cash, or accrual are more useful. This conclusion holds for both the importance and useability dimensions of the construct perceived usefulness.

Over the last two decades, sanitation policy and development has undergone a paradigm shift away from heavily-subsidized, supply-driven approaches towards behavioral-based demand-driven approaches. These current approaches to increase sanitation demand are multi-faceted, requiring multiple stakeholders with varying degrees of interest, knowledge, and capacity. Although efforts exist to increase sanitation access by incorporating engineering design principles with implementation planning approaches, these groups generally work independently without strong connections, thus reducing the potential of their impact. As a result, the design of appropriate sanitation technology is disengaged from the implementation of acceptable technology into communities, disconnecting user preference integration from sanitation technology design and resulting in fewer sanitation technologies being adopted and used. To address these challenges in developing successful interventions, this research examined how user preferences for specific attributes of appropriate sanitation technologies and their respective implementation arrangements influence their adoption and usage. Data for the study included interviews of 1002 sanitation users living in a peri-urban area of South Africa; the surveyed respondents were asked about their existing sanitation technology, their preferences for various sanitation technology design attributes, as well as their perspectives on current and preferred sanitation implementation arrangements. The data revealed that user acceptability of appropriate sanitation technology is influenced by the adoption classification of the users. Through the identification of motives and barriers to sanitation usage that were statistically significant, it exhibited the need to differentiate users who share private sanitation from those use communal sanitation facilities. Results also indicated that user acceptability of appropriate sanitation systems is dependent on the technical design attributes of sanitation. The development of utility functions detailed the significance of seven technical design attributes and determined their respective priorities. An agent-based simulation examined how user preferences for sanitation technology design and implementation influence its adoption and usage. Findings suggest that user acceptability of sanitation technology is dependent on both the technology design and the implementation arrangement being preferred.

This study strives to solicit and assimilate the attitudes and preferences of online users in Thailand concerning their privacy rights. It is designed to resolve this noticeable void in Thai information technology policy. Three sub problems are separately investigated by use of a questionnaire used to discern Thai uneasiness about (1) the clandestine commercial collection and dissemination of personal online data profiles, (2) a wide variety of online criminal activities that should merit government regulation and intervention, and (3) Internet users' trust in their government to curtail online criminal activities. Scholarly research in Thailand has not previously probed or examined these issues. Statistical methods employed in this study involve the use of "sign test" procedures, since no restrictive speculations are assumed about the population distribution. A questionnaire is employed to elicit replies from a database consisting of knowledgeable Thai Internet users. The resultant outcomes are based on a statistical, nationwide sampling plan that draws its responses from computer literate students at four major regionally diverse public universities. The results reveal that discernable and deep-rooted attitudes of Internet users in Thailand toward online privacy rights exist. Further, the results indicate that online patrons have significant preferences about perceived invasions of their privacy on the Internet. The analyses of these discernible traits provide a groundbreaking profile about Internet users in Thailand that merits further investigations in the future. The findings and implications of this study should provide the cornerstone for legislative bodies, telecommunication regulators, policing authorities, and a wide variety of cyberspace consumers to adhere to these concerns, and to enhance the orderly expansion of advanced Internet technologies in Thailand. Ultimately, the resolution of these matters should contribute to the development of a useful IT infrastructure that will provide fairness to both sides of the Internet equation, benefiting both Internet users and online commercial vendors justly.

In this research end user privacy preferences regarding online resources web and mobile applications and websites are investigated and design requirements needed for the development of a privacy focused, privacy enhancing technology tool are identified, as derived from the literature, the crowd source based solution is the most appealing solution so it is fully analyzed according to our research main focus.

By now, people are accustomed to getting some personalized recommendations when they are finding movies to watch, music to listen, and so on. All of these recommendations come from recommender systems, and can aid the process of the decision making to avoid the problem of "information overload". Over the years, there has been much work done both in industry and academia on developing new approaches for recommender systems. However, there are still some hurdles in adapting recommender systems to a broader range of real-life applications. In the e-commerce environment especially with the so called high-risk products (also called high-cost or high-involvement products, such as digital cameras, computers, and cars), because a user does not buy the high-risk product very often, it is normal that s/he is not able to rate many products. For the same reason, the current buyer is often a new user because s/he would not afford to buy the same kind of high-risk product before. The traditional recommender techniques (such as user-based collaborative filtering and content-based methods) can thus not be effectively applicable in this environment, because they largely assume that the users have prior experiences with products. Thus, the "data sparsity" and "new users" are two typical challenging issues that the classical recommender systems cannot well address in high-risk product domains. In some recommender systems, a new user will be asked to indicate his/her preferences on some aspects in order to address the so called cold-start problem via collecting some preferences. Such collected preferences are usually not complete due to unfamilaring with the product domain, which are called partial preferences.;In this thesis, we propose to leverage some auxiliary data of online reviewers' opinions, so as to enrich the partial preferences. With the objective of developing more effective recommender systems for high-risk products in e-commerce, in our work, we have exerted to derive reviewers' preferences from the textual reviews they posted. Then, these recovered preferences are leveraged to estimate and supplement a new buyer's preference with which the product recommendation is produced. Firstly, we propose a novel clustering method based on Latent Class Regression model (LCRM), which is able to consider both the overall ratings and feature-level opinion values (as extracted from textual reviews) to infer individual reviewers' weight feature preferences, that represent the weights the user places on different product features. Secondly, we propose a method to estimate reviewers' value preferences (i.e., the user's preferences on the product's attribute values) by matching their review opinions to the corresponding attributes' static specifications. Thirdly, we investigate how to combine weight preferences and value preferences to model user preferences based on Multi-Attribute Utility Theory (MAUT) with the purpose of providing higher quality product recommendations. Particularly, it was shown from our experimental studies that the incorporation of review information can significantly enhance the recommendation accuracy, relative to those without considering reviews. As the practical implication, our proposed solutions can be usefully plugged into an online system to be adopted in real-ecommerce sites.

Companies across a wide range of business areas are working hard to fulfill users wishes to speak to digital voice assistants. The trend of replacing chatbots in favour for voice assistants carries a risk of companies not considering which applications will actually benefit from getting a voice user interface (VUI) resulting in poor user experience.  This thesis aims to investigate which help desk support task will benefit from being implemented in a VUI. By following the Service Design methodology, research on the topic has been conducted and a prototype has been build and tested on a target audience. The results from a user study were evaluated and conclusions have been drawn about which tasks are best suited for being handled by a digital voice assistant.  Two kinds of help desk tasks were evaluated in a user study to compared users experience of the current text based digital assistant with a prototype of a voice based assistant. The aim of the user study was to find which task would benefit from becoming voice based by looking at users acceptance level and over all experience.  The results from the user study showed that employees who use the current text based assistant for help desk tasks, will not choose to speak to a digital voice assistant because they are happy with the service available today. However, employees who don’t use the current text based assistant, will find the digital voice assistant useful. It was also found that short executing tasks such as unlocking accounts, are a better fit for the VUI compared to longer interactions providing information. Two conclusions were drawn, peoples’ preferences are different, meaning that it should be possible to interact with both a text based and voice based assistant when performing help desk tasks. Secondly, the voice based assistant should be implemented as a function in the help desk phone queue instead of being implemented in a browser. Because the users argued that they would be more comfortable speaking to a phone then to a screen.<br>Företag i alla branscher jobbar hårt för att uppfylla sina användaresönskan om att interagera med röstassistenter. Trenden att byta en chattbot mot en röstassistent medför en risk att företag inte tar hänsyn till huruvida en tjänst faktiskt drar nytta av att göras om till ett röstbaserat användargränssnitt, vilket kan resultera i en försämrad användarupplevelse.  Denna uppsats undersöker vilka funktioner i en kundtjänst som skulle gagnas av att implementeras i ett röstbaserat användargr ̈anssnitt. Genom att använda Service Design modellens forsknings- och idé-genererings fas har en röstbaserad prototyp tagits fram och testats på målgruppen. Resultaten från användarstudien har utvärderats och slutsatser har formulerats.  Två typer av kundtjänstfunktioner har undersökts i en användarstudie som jämfört användarnas upplevelse av den befintliga chatbotten och en röstassistentsprototyp. Målet med användarstudien var att definiera vilka kundtjänstfunktioner som skulle gynnas av att bli röstbaserade genom att titta på användarnas acceptansnivå och övergripande upplevelse.  Resultaten visar att, användare som idag använder, och är nöjda med, chatbotten förmodligen inte kommer att ersätta denna med röstassistenten. Samtidigt som användaren som idag inte använder chatbotten kan tänka sig att använda röstassistenten istället för att ringa till kundtjänsten.  En annan upptäckt från användarstudien var att funktioner som utför en uppgift, så som att låsa upp ett konto, passar bättre i ett röstbaserat sammanhang i jämförelse med när längre information ska förmedlas.  Slutligen formulerades två slutsatser. För det första, olika personer har olika preferenser, det borde alltså vara möjligt att interagera med både chatbotten och röstassistenten för kundtjänstärenden. För det andra, röstassistenten borde implementeras som en plugin som användaren kan utnyttja när denne sitter i telefonkön till kundtjänsten snarare än en egen funktion i på den befintliga hemsida. Detta på grund av att användarna uttryckte att det är mer bekväma med att prata i telefon snarare än till en skärm.

While traditional Saudi Arabian courtship is rigidly structured according to a set of Islamic codes of conduct, over the past decade, Saudis have increasingly turned to unconventional means of finding and courting a potential spouse: matrimonial websites. Given that little research has been done on this emerging form of courtship, which differs substantially from Western-style online dating, this research examines the impressions Saudi users intended to form when constructing their profiles on matrimonial websites, the characteristics these users seek in their potential spouses, and the acceptable Saudi script for courtships initiated on matrimonial websites. It unites hyperpersonal theory, sexual strategies theory, and script theory into a theoretical framework. The research consisted of three sequential phases, starting with a quantitative questionnaire (N = 302), followed by quantitative content analysis of the website profiles (N = 111), and ending with a qualitative semi-structured interview phase (N = 36). The results of the research contribute to the debate in the literature on Computer Meditated Communication regarding whether online settings provide rich information about other users, as they show that online interactions can provide more information about the opposite sex for users who belong to gender-segregated societies than they can get through their offline lives. The results also reveal that there are similarities and differences between Buss and Schmitt’s (1993) proposed strategic mate preferences and Saudis’ preferences. Such findings contribute to the theories on mate preferences in general and sexual strategies theory in particular by advancing the understanding of mate preferences in an Islamic context. The results also provided an in-depth description of the script Saudi users follow in trying to find a potential spouse through matrimonial websites. The findings also contribute to the online dating literature by showing the similarities and differences between conservative, Islamic Saudi users and Western users in using websites to search for a potential mate. Although these Saudi users deviate from traditional methods of finding a potential spouse, the findings of this research reveal that they do not completely challenge their traditions. The research shows the power of both social and religious norms in affecting these users’ behaviours and decisions when using matrimonial websites.

<p> In the decade since 2007, public bikeshare systems (PBS) have disrupted the landscape of urban transportation systems all over the world. The rapid pace at which urban systems are enduring this disruption due to PBS has left cities and researchers play catch up on understanding various factors impacting the usage and impacts of PBS. Comprehending the profiles and preferences of bikeshare users have a substantial role to play in policy-making, planning and operational management at PBS. However, the research is scant related to these factors. </p><p> As its first major objective, this research evaluated the impact of pricing on bikeshare ridership and revenue. As a case study, the introduction of single-trip fare (STF) for $2 by Capital Bikeshare (CaBi) was studied. Aggregate analysis results showed that the first-time casual (short-term) users increased by as much as 79% immediately after the introduction of STF. Jurisdiction-level analysis indicated a statistically significant increase in casual user ridership for identical 12-month periods before and after the introduction of STF. The introduction of STF did not impact ridership and revenues of registered (annual or monthly) members. Casual user revenues before and after the introduction of STF were also compared at the station-level, while controlling seasonal and weather factors. The results showed a statistically significant increase in ridership and decrease in revenue per ride for casual users after the introduction of STF. </p><p> Even though casual bikeshare users account for a large share of revenue, literature provides very little insights about the casual users. As the second major objective of this research, profiles and preferences of bikeshare users (registered members and casual users) were obtained by conducting an intercept survey of CaBi users. Survey findings indicated that, when compared to casual users, registered members are more likely to earn more and are more sensitive to service as reflected by station density. A typical White user has 2.4 times greater odds of being registered members than other race. Analysis also revealed that each additional increase in the number of monthly trips leads to about 18% increase in the odds of the bikeshare user being a registered member. </p><p> As the third major objective, this research evaluated price sensitivities and elasticities of bikeshare fare products using monadic design implemented in the survey instrument. Higher household income groups and White users were found to be less sensitive to price compared to other income groups and other races/ethnicities. Pivot-price elasticities revealed that females are about 30% and 10% more price sensitive than males for single-trip fare (STF) and annual membership, respectively. Also, sightseeing trips are 30% less price sensitive than work trips for STF purchasers. Results from this study would be useful in policy-making, planning and operations for bikeshare systems.</p><p>

This study investigated the effect that aggregation functions and similarity measures had on the accuracy of memory based collaborative filtering algorithms, specifically in the case when the actual rating that a user would provide deviates significantly from the rating that the average user would provide. Nine different combinations of algorithms were implemented by using three different aggregation functions and three different similarity measures. These algorithms were tested on the MovieLens dataset and the similarity measures used were the Pearson correlation, JPSS and JEPSS which is a modified version of JPSS. From the results it was concluded that using the normsum aggregation function lead to a major improvement in accuracy in general. Another conclusion was that using JEPSS over JPSS does not provide a significant increase in accuracy. Finally, the best results for JPSS and JEPSS were obtained when using the least amount of similar users.<br>Denna studie undersökte effekten som aggregeringsfunktioner och likhetsmått hade på noggrannheten av minnesbaserade collaborative filtering algoritmer, speciellt i de fall där användarens rating av ett föremål skiljer sig mycket från den genomsnittliga värdet. Nio olika kombinationer av algoritmer implementerades genom att kombinera tre aggregeringsfunktioner med tre olika likhetsmått. Dessa algoritmer testades på MovieLens databasen och likhetsmåtten som användes var Pearson correlation, JPSS och JEPSS som är en modifierad version av JPSS. Resultaten visar att aggregeringsfunktionen normsum ledde till en stor förbättring i noggrannheten rent allmänt. Dessutom visas att JEPSS inte ger en signifikant förbättring gentemot JPSS. Slutligen, det bästa resultatet för JPSS och JEPSS erhölls när man använder så få användare som möjligt i aggregeringsfunktionen.

En France le traitement du cancer est principalement dominé par les plateaux techniques hospitaliers, mais ceux-ci sont saturés et jugés trop coûteux. Pour ces raisons, les décideurs publics souhaitent développer des structures alternatives : l’hospitalisation à domicile ou les structures de proximités (SP) comme les maisons de santé pluridisciplinaires. Objectif :-Etudier l’adéquation entre les politiques publiques de santé et les préférences des patients et de la population générale.-Tester l’acceptabilité et la validité de la méthode de l’évaluation contingente (EV) (enquête de Disposition A Payer : DAP) comme moyen d’obtenir les préférences en population générale afin de guider la décision publique.Méthode : Nous avons développé un outil d’information et d’aide à la décision informatisé qui contenait 1. L’information sur les trois options de prise en charge : hôpital, domicile, SP 2. Une enquête de mesure des préférences 3. Des questionnaires socio-démographiques et de recueil de l’expérience de soins 4. Pour la population générale uniquement : une EV.Résultats :L’enquête auprès des patients a été réalisée dans 3 établissements (n=386), celle en population générale a été conduite sur un échantillon représentatif de la Région Rhône-Alpes (n=1001). L’hôpital était l’option préférée pour les deux populations 71% des patients contre 42,6% pour la population générale, suivi du domicile avec 24% pour les patients et 38,8% pour la population générale. L’acceptabilité et la validité de l’EC en population générale semble être confirmée par le faible nombre de protestataires et d’outliers ainsi que par la corrélation entre les DAP, le revenu et les préférences<br>In France, cancer treatments are mainly provided in hospitals but they are expensive, crowded and mainly located around big cities. For these reasons, French public health decision makers wish to develop alternative strategies such as home care and proximity structures (typically GPs, nurses and physiotherapists working together in the same medical office). ObjectivesTo study the adequacy between French health policy and patients’ &amp; general population’s preferences. To test the acceptability and validity of the contingent valuation (CV) method as a mean of obtaining values in the general population to guide public decision makers. Methods: We developed a computerized decision aid composed of: 1. Information on the disease management options: hospital, home care and proximity structures; 2. A survey to measure preferences. 3. A socio demographic and an experience of care questionnaire. 4. For the general population only, a CV survey (willingness to pay: WTP) using a payment card was developed Results: The patients survey has been realized in three hospitals(n=386), whereas the general population survey has been conducted in a representative sample of the Rhône-Alps region (n=1001). Both populations preferred hospital, but the proportions are very different. Indeed, hospital is the preferred option for 71% of patients and only for 42,6% of the general population. Moreover, as compared to patients, the general population more often prefer PS, respectively 18,6% and 5 %. The acceptability and validity of the CV method in the general population seemed to be confirmed by the few rates of protesters and outliers, the correlation between preferences, income and WTP

The goal of this master’s thesis was to create a model that predicts preference towards a specific exclusive feature in a subscribed service. It investigated unsupervised and semi-supervised learning to identify customer segments that prefer an specific exclusive feature. These customers segments were then used as targets for supervised learning algorithms to predict which segment a user on the ads-supported version would belong to. Two experiments was preformed, one to investigate and identify customer segments with the help of a survey and secondly, the preference prediction. It was found that Ward’s agglomerative clustering agreed the best with the preference analysis from the survey. Nevertheless, the correlation between the preference survey and the usage clustering was weak. The random forest classifier was preformed the best on the resulting dataset from Ward’s agglomerative clustering. It was concluded that user usage segmentation for the exclusive features showed promising results as well as the over all method. Nevertheless, due to the weak correlation between the survey and the usage clustering it rather predicts usage than preference.<br>Syftet med denna uppsats var att skapa en modell som förutspår preferens för en specifik "exklusiv funktionalitet" i en abonnerad tjänst. I uppsatsen undersöktes unsupervised och semi-supervised learning för att identifiera kundsegment. Dessa kundsegment användes sedan som mål i supervised learning-algoritmer för att förutsäga vilket segment en användare på annons-versionen skulle tillhöra.Två experiment genomfördes, ett för att undersöka och identifiera kundsegment och ett för att undersöka möjligheten till preferensförutsägelse. Det konstaterades att Wards agglomerative clustering-algoritm gav bäst resultat när den jämfördes med preferensanalysen från enkäten. Sambandet mellan preferensundersökningen och användningen av kluster var dock svag. I klassificeringen av det resulterande datasetet från Wards agglomerative clustering-algoritm gav Random forest bäst resultat.Slutsatsen blev att användarsegmenteringen samt den övergripande metoden visade lovande resultat. På grund av den svaga korrelationen mellan enkäten och användarsegmenteringen förutspås enbart an-vändning av de exklusiva funktionerna.

The question of the relationship between, and the importance of usability and aesthetics, in the field of user interface design, has been debated back and forth. It has also been looked at from different perspectives since Raskin (1994) wrote his article on intuitive design. Several experiments have also been conducted over the last twenty years to find out exactly how much each factor matter, what the ultimate user preference is, and if it can be stereotyped. The more complex part of the discussion, however, seems to be the definitions: exactly what is aesthetics, what is usability and how do they affect each other? To find out, I explored the context of these factors from multiple perspectives, to draw the larger conclusions about what affects what. How accurate is the concept of halo when it comes to interface design; can a less aesthetic interface discourage users from exploring its content? Moreover, can a highly usable interface convince its users that the web page is also aesthetically pleasing? In this paper I will explain the ideas of aesthetic and intuitive design based on two fields of study; human computer interaction design and interaction design. That is in the pursuance of understanding user preference and the design decisions behind one of the most popular interfaces on the internet today.

This study is qualitative and interpretive in nature. It examines the perception of 6 people aged 23-32 with decent experience in using the Web on the usability of three user interface website design patterns. These patterns are the ‘hamburger’ icon (an icon used primarily in mobile websites and apps that shows a hidden navigation when clicked), CAPTCHAs (a task that users have to complete to continue browsing a webpage to prevent automated software operating on the webpage) and returning to the homepage. It searches for the characteristics that they desire to see in those three user interface design patterns and the actions that those patterns represent. The participants are reached through interviews and observations and the research pinpoints that although experienced Internet users find the user interface elements relatively usable some usability factors can be worked upon in the chosen design elements and pinpoints what users would want to see changed, the actual changes they want and the problems they actually encounter with the current status of the three (3) design patterns and their usability. More noticeably, the research pinpoints that a “Homepage” button would be more usable than “Home” button which is the de facto standard as of this moment and it shows that the ‘hamburger’ icon is usable enough amongst experienced users, contradicting the research pinpointing that 71 out of 76 fail using the icon (Fichter and Wisniewski, 2016) probably due to the participants’ experience with technology, but other, preferable alternatives to the ‘hamburger’ icon are revealed from the participants which are in line with the current literature. CAPTCHAs are confirmed as a ‘nuisance’ (Pogue, 2012) and the need for CAPTCHAs which are quick to solve emerges which is what forms the perception of usability of the participants.

L’ubiquité des terminaux intelligents équipés de capteurs a donné naissance à un nouveau paradigme de collecte participative des données appelé Crowdsensing. Pour mener à bien les tâches de collecte, divers défis relatifs à l’implication des participants et des demandeurs de services doivent être relevés. Dans ce contexte, nous abordons quatre questions majeures inhérentes à ce problème: Comment affecter les tâches de collecte afin de maximiser la qualité des données d’une façon éco-énergétique ? Comment minimiser le temps nécessaire à la collecte et au traitement des tâches? Comment inciter les participants à dédier une partie de leurs ressources pour la collecte? et Comment protéger la vie privée des participants tout en préservant la qualité des données reportées ? Tout d’abord, nous nous intéressons au fait que les ressources énergétiques des terminaux mobiles restent limitées. Nous introduisons alors des modèles de déploiement de tâches qui visent à maximiser la qualité des données reportées tout en minimisant le coût énergétique global de la collecte. Ainsi, notre première contribution se matérialise en un modèle d’allocation appelé, QEMSS. QEMSS définit des métriques de qualité de données et cherche à les maximiser en se basant sur des heuristiques utilisant la recherche taboue. De plus, afin de rendre le processus d’allocation résultante plus équitable, nous faisons appel à un deuxième algorithme, F-QEMSS, extension de QEMSS. Les deux solutions ont permis d’obtenir des niveaux de qualité de données compétitifs principalement dans les situations défavorables des zones de faible densité ou de ressources limitées. En outre, afin de minimiser le temps moyen de collecte et de traitement des données, une deuxième phase d’allocation distribuée est ajoutée. Plus précisément, nous proposons dans cette deuxième contribution de désigner des participants responsables de déléguer des tâches. Ces derniers prédisent le comportement d’autres utilisateurs en termes de mobilité et de préférences de collecte. Par conséquent, nous développons deux types d’allocation; MATA qui ne tient compte que de la mobilité et P-MATA qui tient compte à la fois de la mobilité et des préférences des participants. Les deux allocations démontrent que l’estimation des préférences des utilisateurs minimise le temps de collecte et évite le rejet des tâches. La troisième contribution de cette thèse, IP-MATA+, propose des incitations aux participants, ce qui favorise leur engagement aux campagnes de collecte notamment quand le budget dédié est partagé en fonction de la qualité des contributions. Pour finir, nous considérons la problématique de la vie privée des participants au crowdsensing. Particulièrement, nous ciblons la minimisation du risque de divulgation de la vie privée durant la phase du déchargement tout en veillant à l’utilité des données collectées. Ainsi, la quatrième contribution de cette thèse vise à assurer simultanément deux objectifs concurrents, à savoir assurer l’utilité des données nécessaire aux demandeurs et protéger les informations sensibles des participants. Pour ce faire, nous introduisons une entité de confiance dans le système de collecte ayant pour rôle d’exécuter un mécanisme qui génère une version altérée de la donnée collectée qui répond au compromis de protection et d’utilité. La solution développée, appelée PRUM, a été évaluée sur des datasets de collecte participative en variant les scénarios d’attaque et de déchargement des données. Les résultats obtenus prouvent qu’une altération limitée des données collectées peut assurer une protection des informations sensibles des participants tout en préservant environ 98% de l’utilité des données obtenue pour les demandeurs. Pour conclure, nos contributions abordent diverses problématiques complémentaires inhérentes à la collecte participative des données ouvrant la voie à des mises en œuvre réelles et facilitant leur déploiement<br>The ubiquity of sensors-equipped mobile devices has enabled people to contribute data via crowdsensing systems. This emergent paradigm comes with various applications. However, new challenges arise given users involvement in data collection process. In this context, we introduce collaborative sensing schemes which tackle four main questions: How to assign sensing tasks to maximize data quality with energy-awareness? How to minimize the processing time of sensing tasks? How to motivate users to dedicate part of their resources to the crowdsensing process ? and How to protect participants privacy and not impact data utility when reporting collected sensory data ? First, we focus on the fact that smart devices are energy-constrained and develop task assignment methods that aim to maximize sensor data quality while minimizing the overall energy consumption of the data harvesting process. The resulting contribution materialized as a Quality and Energy-aware Mobile Sensing Scheme (QEMSS) defines first data quality metrics then models and solves the corresponding optimization problem using a Tabu-Search based heuristic. Moreover, we assess the fairness of the resulted scheduling by introducing F-QEMSS variant. Through extensive simulations, we show that both solutions have achieved competitive data quality levels when compared to concurrent methods especially in situations where the process is facing low dense sensing areas and resources shortcomings. As a second contribution, we propose to distribute the assignment process among participants to minimize the average sensing time and processing overload com- pared to a fully centralized approach. Thus, we suggest to designate some participants to carry extra sensing tasks and delegate them to appropriate neighbors. The new assign- ment is based on predicting users local mobility and sensing preferences. Accordingly, we develop two new greedy-based assignment schemes, one only Mobility-aware (MATA) and the other one accounting for both preferences and mobility (P-MATA), and evaluate their performances. Both MATA and P-MATA consider a voluntary sensing process and show that accounting for users preferences minimize the sensing time. Having showing that, our third contribution in this thesis is conceived as an Incentives-based variant, IP-MATA+. IP-MATA+ incorporates rewards in the users choice model and proves their positive impact on enhancing their commitment especially when the dedicated budget is shared function of contributed data quality. Finally, our fourth and last contribution addresses the seizing of users privacy concerns within crowdsensing systems. More specifically, we study the minimization of the incurred privacy leakage in data uploading phase while accounting for the possible quality regression. That is, we assess simultaneously the two competing goals of ensuring queriers required data utility and protecting participants’ sensitive information. Thus, we introduce a trust entity to the crowdsensing traditional system. This entity runs a general privacy-preserving mechanism to release a distorted version of sensed data that responds to a privacy-utility trade-off. The proposed mechanism, called PRUM, is evaluated on three sensing datasets, different adversary models and two main data uploading scenarios. Results show that a limited distortion on collected data may ensure privacy while maintaining about 98% of the required utility level.The four contributions of this thesis tackle competing issues in crowdsensing which paves the way at facilitating its real implementation and aims at broader deployment

PONTIFÍCIA UNIVERSIDADE CATÓLICA DO RIO DE JANEIRO<br>CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO<br>A escolha de uma entre um conjunto de opções disponível normalmente requer a resolução de trade-offs, contudo esperar que as pessoas avaliem cada uma das opções de um grande conjunto pode ser inviável devido ao tempo e ao esforço cognitivo necessários para realizar tal análise, fazendo com que elas fiquem freqüentemente insatisfeitas com suas escolhas. Sistemas de software podem dar suporte à tomada de decisão humana ou mesmo automatizar esse processo, entretanto existem muitos desafios que estão associados com o oferecimento de tal suporte. Esta tese lida, em particular, com três destes desafios: (i) como representar preferências dos usuários; (ii) como raciocinar sobre estas preferências e tomar decisões; e (iii) como justificar tais decisões. Diferentes abordagens têm sido propostas para a representação e raciocínio sobre preferências qualitativas, mas estas abordagens lidam com um conjunto restrito de tipos de preferências, e portanto não são capazes de processar preferências fornecidas por usuários em muitos cenários realistas. Nesta tese, apresentam-se três principais contribuições. A primeira delas consiste de um novo metamodelo de preferências, o qual foi desenvolvido de acordo com um estudo sobre a expressão de preferências, permitindo a representação de preferências em alto-nível. Segundo, uma nova técnica de tomada de decisão automatizada é proposta, a qual escolhe uma opção de um conjunto de opções disponível baseada em preferências expressas em uma linguagem construída de acordo com o metamodelo proposto, explorando termos da linguagem natural, tais como atos de fala expressivos. A técnica vai além das preferências fornecidas para tomar a decisão através da incorporação de princípios da psicologia, que focam como os humanos tomam decisões, já que as preferências fornecidas tipicamente não são suficientes para resolver trade-offs entre as opções disponíveis. Terceiro, apresenta-se uma técnica de geração de explicação, que utiliza modelos construídos pela técnica de tomada de decisão para justificar escolhas, e segue diretrizes e padrões que foram derivados de um estudo sobre explicações a respeito de escolhas, também realizado no contexto desta tese. Um estudo com usuários foi feito para avaliar a abordagem, o qual mostra que (i) a linguagem de preferências é adequada para usuários expressarem suas preferências, que (ii) a técnica de tomada de decisão faz escolhas que os usuários consideram de alta qualidade, e que (iii) as explicações fornecidas permitem que usuários entendam por que a escolha foi feita, bem como melhora a confiança na decisão tomada.<br>Choosing from a set of available options often requires resolution of trade-offs but it can be unfeasible for humans to carefully evaluate each option of a large set due to the required time and cognitive effort. Consequently, they are often unsatisfied with their choices. Software systems can support human decision making or even automate this process, but there are many challenges associated with the provision of such support. In this thesis we deal in particular with three of them: (i) how to represent user preferences; (ii) how to reason about preferences and make decisions; and (iii) how to justify such decisions. Different approaches have been proposed for representing and reasoning about qualitative preferences, but they address a restricted set of preference types, and therefore are not able to process preferences provided by users in many realistic scenarios. This thesis provides three main contributions. First, we introduce a new preference metamodel founded on a study of how humans express preferences, allowing the representation of high-level preferences. Second, we propose an automated decision making technique, which chooses an option from a set available based on preferences expressed in a language based on our metamodel, exploiting natural-language terms. Our technique goes beyond the provided preferences to make a decision with the incorporation of psychology principles, which concern how humans make decisions, as the provided preferences are typically not enough to resolve trade-offs among available options. Third, we present an explanation generation technique, which uses models built by our decision making technique to justify choices, and follows guidelines and patterns that we derived from a study of choice explanation. A user study was performed to evaluate our approach, which shows that (i) our preference language is adequate for users to express their preferences, (ii) our decision making technique makes choices that users consider as having good quality, and (iii) the provided explanations allow users to understand why the choice was made and improves the confidence in the decision.

Intelligent systems that assist users in fulfilling complex tasks need a concise and processable representation of incomplete and uncertain information. In order to be able to choose among different options, these systems also need a compact and processable representation of the concept of preference. Preferences can provide an effective way to choose the best solutions to a given problem. These solutions can represent the most plausible states of the world when we model incomplete information, the most satisfactory states of the world when we express user preferences, or optimal decisions when we make decisions under uncertainty. Several domains, such as, reasoning under incomplete and uncertain information, user preference modeling, and qualitative decision making under uncertainty, have benefited from advances on preference representation. In the literature, several symbolic approaches of nonclassical reasoning have been proposed. Among them, logic programming under answer set semantics offers a good compromise between symbolic representation and computation of knowledge and several extensions for handling preferences. Nevertheless, there are still some open issues to be considered in logic programming. In nonmonotonic reasoning, first, most approaches assume that exceptions to logic program rules are already specified. However, sometimes, it is possible to consider implicit preferences based on the specificity of the rules to handle incomplete information. Secondly, the joint handling of exceptions and uncertainty has received little attention: when information is uncertain, the selection of default rules can be a matter of explicit preferences and uncertainty. In user preference modeling, although existing logic programming specifications allow to express user preferences which depend both on incomplete and contextual information, in some applications, some preferences in some context may be more important than others. Furthermore, more complex preference expressions need to be supported. In qualitative decision making under uncertainty, existing logic programming-based methodologies for making decisions seem to lack a satisfactory handling of preferences and uncertainty. The aim of this dissertation is twofold: 1) to tackle the role played by preferences in logic programming from different perspectives, and 2) to contribute to this novel field by proposing several frameworks and methods able to address the above issues. To this end, we will first show how preferences can be used to select default rules in logic programs in an implicit and explicit way. In particular, we propose (i) a method for selecting logic program rules based on specificity, and (ii) a framework for selecting uncertain default rules based on explicit preferences and the certainty of the rules. Then, we will see how user preferences can be modeled and processed in terms of a logic program (iii) in order to manage user profiles in a context-aware system and (iv) in order to propose a framework for the specification of nested (non-flat) preference expressions. Finally, in the attempt to bridge the gap between logic programming and qualitative decision under uncertainty, (v) we propose a classical- and a possibilistic-based logic programming methodology to compute an optimal decision when uncertainty and preferences are matters of degrees.<br>Els sistemes intel.ligents que assisteixen a usuaris en la realització de tasques complexes necessiten una representació concisa i formal de la informació que permeti un raonament nomonòton en condicions d’incertesa. Per a poder escollir entre les diferents opcions, aquests sistemes solen necessitar una representació del concepte de preferència. Les preferències poden proporcionar una manera efectiva de triar entre les millors solucions a un problema. Aquestes solucions poden representar els estats del món més plausibles quan es tracta de modelar informació incompleta, els estats del món més satisfactori quan expressem preferències de l’usuari, o decisions òptimes quan estem parlant de presa de decisió incorporant incertesa. L’ús de les preferències ha beneficiat diferents dominis, com, el raonament en presència d’informació incompleta i incerta, el modelat de preferències d’usuari, i la presa de decisió sota incertesa. En la literatura, s’hi troben diferents aproximacions al raonament no clàssic basades en una representació simbòlica de la informació. Entre elles, l’enfocament de programació lògica, utilitzant la semàntica de answer set, ofereix una bona aproximació entre representació i processament simbòlic del coneixement, i diferents extensions per gestionar les preferències. No obstant això, en programació lògica es poden identificar diferents problemes pel que fa a la gestió de les preferències. Per exemple, en la majoria d’enfocaments de raonament no-monòton s’assumeix que les excepcions a default rules d’un programa lògic ja estan expressades. Però de vegades es poden considerar preferències implícites basades en l’especificitat de les regles per gestionar la informació incompleta. A més, quan la informació és també incerta, la selecció de default rules pot dependre de preferències explícites i de la incertesa. En el modelatge de preferències del usuari, encara que els formalismes existents basats en programació lògica permetin expressar preferències que depenen d’informació contextual i incompleta, en algunes aplicacions, donat un context, algunes preferències poden ser més importants que unes altres. Per tant, resulta d’interès un llenguatge que permeti capturar preferències més complexes. En la presa de decisions sota incertesa, les metodologies basades en programació lògica creades fins ara no ofereixen una solució del tot satisfactòria pel que fa a la gestió de les preferències i la incertesa. L’objectiu d’aquesta tesi és doble: 1) estudiar el paper de les preferències en la programació lògica des de diferents perspectives, i 2) contribuir a aquesta jove àrea d’investigació proposant diferents marcs teòrics i mètodes per abordar els problemes anteriorment citats. Per a aquest propòsit veurem com les preferències es poden utilitzar de manera implícita i explícita per a la selecció de default rules proposant: (i) un mètode basat en l’especificitat de les regles, que permeti seleccionar regles en un programa lògic; (ii) un marc teòric per a la selecció de default rules incertes basat en preferències explícites i la incertesa de les regles. També veurem com les preferències de l’usuari poden ser modelades i processades usant un enfocament de programació lògica (iii) que suporti la creació d’un mecanisme de gestió dels perfils dels usuaris en un sistema amb reconeixement del context; (iv) que permeti proposar un marc teòric capaç d’expressar preferències amb fòrmules imbricades. Per últim, amb l’objectiu de disminuir la distància entre programació lògica i la presa de decisió amb incertesa proposem (v) una metodologia basada en programació lògica clàssica i en una extensió de la programació lògica que incorpora lògica possibilística per modelar un problema de presa de decisions i per inferir una decisió òptima.<br>Los sistemas inteligentes que asisten a usuarios en tareas complejas necesitan una representación concisa y procesable de la información que permita un razonamiento nomonótono e incierto. Para poder escoger entre las diferentes opciones, estos sistemas suelen necesitar una representación del concepto de preferencia. Las preferencias pueden proporcionar una manera efectiva para elegir entre las mejores soluciones a un problema. Dichas soluciones pueden representar los estados del mundo más plausibles cuando hablamos de representación de información incompleta, los estados del mundo más satisfactorios cuando hablamos de preferencias del usuario, o decisiones óptimas cuando estamos hablando de toma de decisión con incertidumbre. El uso de las preferencias ha beneficiado diferentes dominios, como, razonamiento en presencia de información incompleta e incierta, modelado de preferencias de usuario, y toma de decisión con incertidumbre. En la literatura, distintos enfoques simbólicos de razonamiento no clásico han sido creados. Entre ellos, la programación lógica con la semántica de answer set ofrece un buen acercamiento entre representación y procesamiento simbólico del conocimiento, y diferentes extensiones para manejar las preferencias. Sin embargo, en programación lógica se pueden identificar diferentes problemas con respecto al manejo de las preferencias. Por ejemplo, en la mayoría de enfoques de razonamiento no-monótono se asume que las excepciones a default rules de un programa lógico ya están expresadas. Pero, a veces se pueden considerar preferencias implícitas basadas en la especificidad de las reglas para manejar la información incompleta. Además, cuando la información es también incierta, la selección de default rules pueden depender de preferencias explícitas y de la incertidumbre. En el modelado de preferencias, aunque los formalismos existentes basados en programación lógica permitan expresar preferencias que dependen de información contextual e incompleta, in algunas aplicaciones, algunas preferencias en un contexto puede ser más importantes que otras. Por lo tanto, un lenguaje que permita capturar preferencias más complejas es deseable. En la toma de decisiones con incertidumbre, las metodologías basadas en programación lógica creadas hasta ahora no ofrecen una solución del todo satisfactoria al manejo de las preferencias y la incertidumbre. El objectivo de esta tesis es doble: 1) estudiar el rol de las preferencias en programación lógica desde diferentes perspectivas, y 2) contribuir a esta joven área de investigación proponiendo diferentes marcos teóricos y métodos para abordar los problemas anteriormente citados. Para este propósito veremos como las preferencias pueden ser usadas de manera implícita y explícita para la selección de default rules proponiendo: (i) un método para seleccionar reglas en un programa basado en la especificad de las reglas; (ii) un marco teórico para la selección de default rules basado en preferencias explícitas y incertidumbre. También veremos como las preferencias del usuario pueden ser modeladas y procesadas usando un enfoque de programación lógica (iii) para crear un mecanismo de manejo de los perfiles de los usuarios en un sistema con reconocimiento del contexto; (iv) para crear un marco teórico capaz de expresar preferencias con formulas anidadas. Por último, con el objetivo de disminuir la distancia entre programación lógica y la toma de decisión con incertidumbre proponemos (v) una metodología para modelar un problema de toma de decisiones y para inferir una decisión óptima usando un enfoque de programación lógica clásica y uno de programación lógica extendida con lógica posibilística.<br>Sistemi intelligenti, destinati a fornire supporto agli utenti in processi decisionali complessi, richiedono una rappresentazione dell’informazione concisa, formale e che permetta di ragionare in maniera non monotona e incerta. Per poter scegliere tra le diverse opzioni, tali sistemi hanno bisogno di disporre di una rappresentazione del concetto di preferenza altrettanto concisa e formale. Le preferenze offrono una maniera efficace per scegliere le miglior soluzioni di un problema. Tali soluzioni possono rappresentare gli stati del mondo più credibili quando si tratta di ragionamento non monotono, gli stati del mondo più soddisfacenti quando si tratta delle preferenze degli utenti, o le decisioni migliori quando prendiamo una decisione in condizioni di incertezza. Diversi domini come ad esempio il ragionamento non monotono e incerto, la strutturazione del profilo utente, e i modelli di decisione in condizioni d’incertezza hanno tratto beneficio dalla rappresentazione delle preferenze. Nella bibliografia disponibile si possono incontrare diversi approcci simbolici al ragionamento non classico. Tra questi, la programmazione logica con answer set semantics offre un buon compromesso tra rappresentazione simbolica e processamento dell’informazione, e diversi estensioni per la gestione delle preferenze sono state proposti in tal senso. Nonostante ció, nella programmazione logica esistono ancora delle problematiche aperte. Prima di tutto, nella maggior parte degli approcci al ragionamento non monotono, si suppone che nel programma le eccezioni alle regole siano già specificate. Tuttavia, a volte per trattare l’informazione incompleta è possibile prendere in considerazione preferenze implicite basate sulla specificità delle regole. In secondo luogo, la gestione congiunta di eccezioni e incertezza ha avuto scarsa attenzione: quando l’informazione è incerta, la scelta di default rule può essere una questione di preferenze esplicite e d’incertezza allo stesso tempo. Nella creazione di preferenze dell’utente, anche se le specifiche di programmazione logica esistenti permettono di esprimere preferenze che dipendono sia da un’informazione incompleta che da una contestuale, in alcune applicazioni talune preferenze possono essere più importanti di altre, o espressioni più complesse devono essere supportate. In un processo decisionale con incertezza, le metodologie basate sulla programmazione logica viste sinora, non offrono una gestione soddisfacente delle preferenze e dell’incertezza. Lo scopo di questa dissertazione è doppio: 1) chiarire il ruolo che le preferenze giocano nella programmazione logica da diverse prospettive e 2) contribuire proponendo in questo nuovo settore di ricerca, diversi framework e metodi in grado di affrontare le citate problematiche. Per prima cosa, dimostreremo come le preferenze possono essere usate per selezionare default rule in un programma in maniera implicita ed esplicita. In particolare proporremo: (i) un metodo per la selezione delle regole di un programma logico basato sulla specificità dell’informazione; (ii) un framework per la selezione di default rule basato sulle preferenze esplicite e sull’incertezza associata alle regole del programma. Poi, vedremo come le preferenze degli utenti possono essere modellate attraverso un programma logico, (iii) per creare il profilo dell’utente in un sistema context-aware, e (iv) per proporre un framework che supporti la definizione di preferenze complesse. Infine, per colmare le lacune in programmazione logica applicata a un processo di decisione con incertezza (v) proporremo una metodologia basata sulla programmazione logica classica e una metodologia basata su un’estensione della programmazione logica con logica possibilistica.

Capturing user preferences is a problematic task. Simply asking the users what they want is too intrusive and prone to error, yet monitoring behaviour unobtrusively and finding meaningful patterns is both difficult and computationally time consuming. Capturing accurate user preferences is, however, an essential task if the information systems of tomorrow are to respond dynamically to the changing needs of their users. This thesis tests the hypothesis that using an ontology to represent user profiles offers advantages over traditional profile representations in the context of recommender systems. A novel ontology-based approach to recommendation is applied to a real world problem and empirically evaluated. Synergy between recommender systems and ontologies is then explored to help overcome both the recommender system cold-start problem and the ontology interest-acquisition problem. Finally, the visualization of profiles in ontological terms is examined in a real world situation and empirically evaluated.

Implicit feedback is usually data that comes from users’ clicks, search queries and text highlights. It exists in abun- dance, but it is riddled with much noise and requires advanced algorithms to properly make good use of it. Several findings suggest that factors such as click-through data and reading time could be used to create user behaviour models in order to predict the users’ information need. This Master’s thesis aims to use click-through data and search queries together with heuristics to create a model that prioritises metadata-fields of the documents in order to predict the information need of a user. Simply put, implicit feedback will be used to improve the precision of a search engine. The Master’s thesis was carried out at Findwise AB - a search engine consultancy firm. Documents from the benchmark dataset INEX were indexed into a search engine. Two different heuristics were proposed that increment the priority of different metadata-fields based on the users’ search queries and clicks. It was assumed that the heuristics would be able to change the listing order of the search results. Evaluations were carried out for the two heuristics and the unmodified search engine was used as the baseline for the experiment. The evaluations were based on simulating a user that searches queries and clicks on documents. The queries and documents, with manually tagged relevance, used in the evaluation came from a data set given by INEX. It was expected that listing order would change in a way that was favourable for the user; the top-ranking results would be documents that truly were in the interest of the user. The evaluations revealed that the behaviour of the heuristics and the baseline have erratic behaviours and metrics never converged to any specific mean-relevance. A statistical test revealed that there is no difference in accuracy between the heuristics and the baseline. These results mean that the proposed heuristics do not improve the precision of the search engine and several factors, such as the indexing of too redundant metadata, could have been responsible for this outcome.<br>Implicit feedback är oftast data som kommer från användarnas klick, sökfrågor och textmarkeringar. Denna data finns i överflöd, men har för mycket brus och kräver avancerade algoritmer för att man ska kunna dra nytta av den. Flera rön föreslår att faktorer som klickdata och läsningstid kan användas för att skapa beteendemodeller för att förutse användarens informationsbehov. Detta examensarbete ämnar att använda klickdata och sökfrågor tillsammans med heuristiker för att skapa en modell som prioriterar metadata-fält i dokument så att användarens informationsbehov kan förutses. Alltså ska implicit feedback användas för att förbättra en sökmotors precision. Examensarbetet utfördes hos Findwise AB - en konsultfirma som specialiserar sig på söklösningar. Dokument från utvärderingsdatamängden INEX indexerades i en sökmotor. Två olika heuristiker skapades för att ändra prioriteten av metadata-fälten utifrån användarnas sök- och klickdata. Det antogs att heuristikerna skulle kunna förändra ordningen av sökresultaten. Evalueringar utfördes för båda heuristiker och den omodifierade sökmotorn användes som måttstock för experimentet. Evalueringarna gick ut på att simulera en användare som söker på frågor och klickar på dokument. Dessa frågor och dokument, med manuellt taggad relevansdata, kom från en datamängd som tillhandahölls av INEX. Evalueringarna visade att beteendet av heuristikerna och måttstocket är slumpmässiga och oberäkneliga. Ingen av heuristikerna konvergerar mot någon specifik medelrelevans. Ett statistiskt test visar att det inte är någon signifikant skillnad på uppmätt träffsäkerhet mellan heuristikerna och måttstocket. Dessa resultat innebär att heuristikerna inte förbättrar sökmotorns precision. Detta utfall kan bero på flera faktorer som t.ex. indexering av överflödig meta-data.

En raison du grand succès des réseaux sans _l et des appareils portatifs, le paradigme de l'informatique pervasive est devenu une réalité. L'un des plus di_ciles objectifs à atteindre dans de tels environnements est de permettre à l'utilisateur d'exécuter une tâche en composant à la volée, les services et les ressources de l'environnement. Cela implique la correspondance et la sélection automatique de services à travers divers dispositifs de l'environnement pervasif. Les approches existantes considèrent souvent seulement les aspects fonctionnels des services et ne prennent pas en compte diff érents aspects non-fonctionnels tels que les préférences utilisateur, les capacités des dispositifs en termes matériels et logiciels, et l'hétérogénéité du réseau de ces dispositifs. Nous présentons une approche pour la sélection dynamique des composants et des dispositifs dans un environnement pervasif en considérant simultanément tous les aspects précédemment mentionnés. Premièrement, nous proposons une modélisation abstraite et concrète de l'application, des capacités des terminaux et des ressources, des préférences des utilisateurs, ainsi que la modélisation de la plate-forme réseau sous-jacente. Les capacit és des dispositifs sont représentées par notre extension du modèle CC/PP et les préférences des utilisateurs en utilisant notre extension du modèle CP-Net. Nous mod- élisons sous forme d'un graphe la tâche de l'utilisateur et des services réseau sous-jacent, ainsi que les exigences des services, des préférences utilisateur et les capacités des dispositifs. L'hétérogénéité des protocoles de communication est également considérée dans les graphes. Les aspects algorithmiques ont été traités en fournissant des algorithmes pour la correspondance entre les services et les composants, pour la projection des applications sur la plate-forme de composants existants et pour l'évaluation des préférences utilisateurs. Pour la description de la composition de l'application nous proposons un modèle SCA étendu. Partant d'une composition abstraite de services, nous arrivons à réaliser une composition concrète de l'application distribuée à travers les dispositifs existants. Si pendant l'exécution un nouveau meilleur dispositif apparaît, l'application est recomposée en tenant compte des nouveaux composants. Cela permet de réaliser la continuité de la session d'un dispositif vers un autre. Une mise en oeuvre d'un prototype et son évaluation sont également fournis<br>Due to the large success of wireless networks and portable devices, the pervasive computing paradigm is becoming a reality. One of the most challenging objectives to be achieved in pervasive computing environments is to allow a user to perform a task by composing on the _y the environment's service and resource components. This involves automatic matching and selection of services across various devices in the pervasive environment. Existing approaches mostly consider only functional aspects for service and component matching and do not consider various non-functional aspects such as user preferences, device capabilities in terms of software and hardware, and network heterogeneity of devices. We present an approach for dynamic selection of components and devices in a pervasive environments considering all the aforementioned aspects simultaneously. First, we provide a modeling of abstract and concrete application, device capabilities and resources, user preferences as well as modeling of the underlying connected platform. Device capabilities are represented by our extended CC/PP model and user preferences using our extended CP-net model. We model both the user task and the underlying network services, along with service requirements, user preferences and device capabilities, as graphs. The heterogeneity of communication protocols is also considered in the graph. The algorithmic aspects have been treated by providing algorithms for service and component matching, application mapping on network platform and user preference evaluation. For description of application composition extended SCA model is used. Departing from an abstract composition, we arrive on achieving a concrete application composition which may be distributed across more than one device. If during the application execution a new, better device appears, the application is recomposed to replace the existing components by the newer ones. This also implies the continuity of session from one device to another. A prototype implementation and its evaluation are also provided

In the current study, it was aimed to explore young adults&rsquo<br>communication technology preferences in friendships. With this aim, two survey studies were conducted on university students. In the first study 178 students and in the second study 343 students were surveyed. The findings of both studies demonstrated that (a) face-to-face communication was the most preferred communication,(b) preference for face-to-face communication was positively associated to preference for auditory communication, and negatively associated to preference for written communication, (c) several individual difference variables contributed to relative preferences for face-to-face over mediated communication in negative situations. General &lsquo<br>ndividual difference variables in relation to Relative Preferences were found to fit into a two-factor higher-order structure. The higher order constructs were Social Openness (the factors tapping to individuals&rsquo<br>motivation to engage in interpersonal interactions even when those interactions are challenging) and Individuated Functioning (the factors tapping to individuals&rsquo<br>motivation for personal growth). Relative Preference increased with increases in Social Openness and in Individuated Functioning. Communicaiton-specific individual difference factors underling the associations among Social Openness, Individauted Functioning and Relative Preferences were explored. Results suggested that Social Openness &ndash<br>but not Individuated Functioning- was indirectly linked to Relative Preferences via Perceived Difficulty of Communication Situations (Study1)<br>Social Openness was indirectly, and Individuated Functioning was both directly and indirectly linked to Relative Preferences via Difficulty in Handling Negative Situations (Study 2). Theoretical and practical implications of studies were discussed considering the previous literature.

Robotic assistants have inspired numerous books and science fiction movies. In the real world, these kinds of devices are a growing need in amongst the elderly, who while life continue requiring more assistance. While life expectancy is increasing, life quality is not necessarily doing so. Thus, we may find ourselves and our loved ones being dependent and needing another person to perform the most basic tasks, which has a strong psychological impact. Accordingly, assistive robots may be the definitive tool to give more quality of life by empowering dependent people and extending their independent living. Assisting users to perform daily activities requires adapting to them and their needs, as they might not be able to adapt to the robot. This thesis tackles adaptation and personalization issues through user preferences. We 'focus on physical tasks that involve close contact, as these present interesting challenges, and are of great importance for he user. Therefore, three tasks are mainly used throughout the thesis: assistive feeding, shoe fitting, and jacket dressing. We first describe a framework for robot behavior adaptation that illustrates how robots should be personalized for and by end- users or their assistants. Using this framework, non-technical users determine how !he robot should behave. Then, we define the concept of preference for assistive robotics scenarios and establish a taxonomy, which includes hierarchies and groups of preferences, grounding definitions and concepts. We then show how the preferences in the taxonomy are used with Al planning systems to adapt the robot behavior to the preferences of the user obtained from simple questions. Our algorithms allow for long-term adaptations as well as to cope with misinformed user models. We further integrate the methods with low-level motion primitives that provide a more robust adaptation and behavior while lowering the number of needed actions and demonstrations. Moreover, we perform a deeper analysis in Planning and preferences with the introduction of new algorithms to provide preference suggestions in planning domains. The thesis then concludes with a user study that evaluates the use of the preferences in the three real assistive robotics scenarios. The experiments show a clear understanding of the preferences of users, who were able to assess the impact of their preferences on the behavior of the robot. In summary, we provide tools and algorithms to design the robotic assistants of the future. Assistants that should be able to adapt to the assisted user needs and preferences, just as human assistants do nowadays.<br>Els assistents robòtics han inspirat nombrosos llibres i pel·lícules de ciència-ficció al llarg de la història. Però tornant al món real, aquest tipus de dispositius s'estan tornant una necessitat per a una societat que envelleix a un ritme ràpid i que, per tant, requerirà més i més assistència. Mentre l'esperança de vida augmenta, la qualitat de vida no necessàriament ho fa. Per tant, ens podem trobar a nosaltres mateixos i als nostres estimats en una situació de dependència, necessitant una altra persona per poder fer les tasques més bàsiques, cosa que té un gran impacte psicològic. En conseqüència, els robots assistencials poden ser l'eina definitiva per proporcionar una millor qualitat de vida empoderant els usuaris i allargant la seva capacitat de viure independentment. L'assistència a persones per realitzar tasques diàries requereix adaptar-se a elles i les seves necessitats, donat que aquests usuaris no poden adaptar-se al robot. En aquesta tesi, abordem el problema de l'adaptació i la personalització d'un robot mitjançant preferències de l'usuari. Ens centrem en tasques físiques, que involucren contacte amb la persona, per les seves dificultats i importància per a l'usuari. Per aquest motiu, la tesi utilitzarà principalment tres tasques com a exemple: donar menjar, posar una sabata i vestir una jaqueta. Comencem definint un marc (framework) per a la personalització del comportament del robot que defineix com s'han de personalitzar els robots per usuaris i pels seus assistents. Amb aquest marc, usuaris sense coneixements tècnics són capaços de definir com s'ha de comportar el robot. Posteriorment definim el concepte de preferència per a robots assistencials i establim una taxonomia que inclou jerarquies i grups de preferències, els quals fonamenten les definicions i conceptes. Després mostrem com les preferències de la taxonomia s'utilitzen amb sistemes planificadors amb IA per adaptar el comportament del robot a les preferències de l'usuari, que s'obtenen mitjançant preguntes simples. Els nostres algorismes permeten l'adaptació a llarg termini, així com fer front a models d'usuari mal inferits. Aquests mètodes són integrats amb primitives a baix nivell que proporcionen una adaptació i comportament més robusts a la mateixa vegada que disminueixen el nombre d'accions i demostracions necessàries. També fem una anàlisi més profunda de l'ús de les preferències amb planificadors amb la introducció de nous algorismes per fer suggeriments de preferències en dominis de planificació. La tesi conclou amb un estudi amb usuaris que avalua l'ús de les preferències en les tres tasques assistencials. Els experiments demostren un clar enteniment de les preferències per part dels usuaris, que van ser capaços de discernir quan les seves preferències eren utilitzades. En resum, proporcionem eines i algorismes per dissenyar els assistents robòtics del futur. Uns assistents que haurien de ser capaços d'adaptar-se a les preferències i necessitats de l'usuari que assisteixen, tal com els assistents humans fan avui en dia.

Migration is becoming a major political and social issue in a global context. However, while immigrants’ integration into the hosting society and cities has sparked research investigations, there have been few studies focusing on how their presence affects the urban fabric, especially their use of public open spaces and even less is known about how, and in what way, such spaces have an impact on the immigrants themselves. The scope of this research is to gain a better understanding of how immigrants use, perceive and experience public open spaces. The research focuses on the case study of public open spaces in Lisbon, the capital of Portugal. The thesis centres on the experience of first generation immigrants from the three largest immigrant communities living in Portugal, namely, from Brazil, Cape Verde and Ukraine. These immigrants not only represent different waves of migration but also communities from very different geographic origins and socio-cultural backgrounds. David Canter’s ‘Theory of Place’ (Canter, 1977) is used as the theoretical framework for the study, whereby place is understood to be the locus for the juxtaposition of activities that people engage in, their perceptions of it and the physical attributes of space. The key research aim is to explore immigrants’ attitudes, preferences, perceptions, uses and ‘place attachment’ in relation to public open spaces in Lisbon. A mixed-method approach is used to gather information from these three immigrant communities and to establish the relationship between them and public open spaces in the host country. The qualitative methods comprise focus-group discussions and ‘go-along’ interviews, while the quantitative methods include questionnaires. The key findings from the analysis of the cross-cultural experiences show that ‘frequency of use’ is likely to be affected by immigrants’ national and cultural identity. The analysis also reveals the particular meaning that ‘being close to water’ has for the three immigrant groups, especially in terms of its connection with the sea and going to the beach, as well as the importance to them of music being played outdoors and their need for more places to have barbecues. The findings also highlight the importance of public open spaces to immigrants’ lives, particularly in terms of how they evoke different memories (childhood and adult) and how some immigrants have already developed memories in relation to certain places and features in Lisbon. ‘Place attachment’, thus, is shown to serve as an anchor for people and it creates links between the homeland and host country, giving a sense of continuity to immigrants’ lives.

The aim of this thesis has been to theoretically investigate how a context- aware system for mobile applications should behave when a phone call is received while driving in a general setting, while the system considers user preferences and norms. By norm, we mean a sort of rule that a user has to follow in a given environment, e.q. traffic. The investigation has lead to an approach on how a general setting could be implemented; in particular, by general setting we mean that the user is a car driver and the environment is the public roads. A given system (e.q. an app) supports the interaction between the user and the environment. A use case has been used as a basis for the investigation, that involves forwarding calls while considering user preferences and norms while driving in front of a school. The first steps, in this thesis, have been to model the user profile and the environment profile, and to find activation and deactivation conditions of both user preferences and norms. An algorithm in pseudocode has been provided that manages how a context-aware system will handle the context-aware decision making process. The decision making process has then been evaluated, and the evaluation has been done while considering real mobile applications on the market. The results of this thesis can be used, for example, as a basis for implementing context-aware mobile applications, e.q. applications handling calls.

Social hållbarhet är ett ämne som är mer aktuellt nu än någonsin, även om det sällan diskuteras i samband med estetik i den bebyggda miljön. De miljöer vi bygger är också de platser där vi tillbringar majoriteten av vår tid, och de påverkar hur vi mår. I takt med att städer växer och blir tätare så blir gestaltningen av dessa platser allt viktigare. Estetiken utgör en stor del av den bebyggda miljön och vår upplevelse när vi rör oss i städerna. Studier visar på att människor föredrar vissa estetiska element framför andra, och att både kropp och hjärna reagerar på estetiska stimuli. I den här uppsatsen utforskas sambandet mellan vad den forskning som finns kring exteriör byggnadsestetik säger, och hur estetiska preferenser inkluderas och diskuteras i stadsbyggnadsprojektet Rosendal i Uppsala i Sverige.<br>Social sustainability is discussed more today than ever before, but seldom in the context of aesthetics in the built environment. The environments we build are where we spend most of our time, and they affect our wellbeing. With cities growing faster and becoming denser, how we design our home becomes more and more important. Aesthetic variables are a large part of what the built environment is, and constitute much of what we experience on a daily basis when moving around our city. Studies suggest that some aesthetic elements are preferred over others, and that both body and brain respond to aesthetic stimuli. This thesis explores the relationship between what has been found on the subject of building exterior aesthetics so far, and how evidence on aesthetic preferences is addressed in the development project of Rosendal in Uppsala, Sweden

This thesis analyzes the relationship between users&rsquo<br>characteristics and users&rsquo<br>interface preferences. An online survey is developed for this study. This survey composed of two types of questions: (1) users&rsquo<br>personal information such as age, gender, country, cognitive structure, and also computer experience and (2) user interface elements. More than 2,500 participants from 120 different countries throughout the world completed our survey. Results were analyzed using cross tables. Our findings show that there is a relationship between users&rsquo<br>characteristics and users&rsquo<br>interface preferences. In the presence of this relationship, an artificial neural network model is developed for the estimation of the interface preferences based on the user characteristics.

Predicting user preferences is a common problem for many companies and services. With the growth of Internet services it becomes both more important and more lucrative being able to predict what products a user would like and then recommend these to them. There are many ways of attempting this, but this study attempts to use random indexing to solve the same problem. Random indexing is a method that has been used successfully when studying the similarity between words, and allows entities to be represented as vectors with relatively small dimensionality. This would allow for fast and memory-efficient implementations of prediction systems. This study uses the dataset Amazon Fine Food Reviews, which contains reviews of products with a rating. It is attempted to predict these ratings, and the result of random indexing is compared to the results on the same dataset using collaborative filtering. Various parameters used in the random indexing method are also varied, to study their effect on the results. These methods are evaluated based on root mean square error and mean absolute error. The results indicate that random indexing does not generate as good results as collaborative filtering. However, the difference is small enough to warrant further study into the other strengths of random indexing, such as speed and memory efficiency. It is theorized that the sparsity of the dataset might have caused the differences in errors between the methods, and with a dense dataset the results might be better.

In an increasingly populated area of mobile applications craving a user’s attention it is of increasing importance to engage the user through a personalized user experience. This includes delivering content which the user is likely to enjoy as well as showcasing that content in a way such that the user is likely to interact with it. This thesis details the exploration of personalizing the content density of articles in the popular mobile application Flipboard. We use past user behavior to power classification of content density preference using random forests. Our results indicate that a personalized presentation of content does increase the overall engagement of Flipboard’s users however the error rate is current too high for the classifier to be usefu.l

Automated systems for large-span roof truss optimization provide engineers with the flexibility to consider multiple alternatives during conceptual design. This investigation extends previous work on multi-objective roof truss optimization to include the design preferences of a human user. The incorporation of user preferences into the optimization process required creation of a mechanism to identify and model preferences as well as discovery of an appropriate location within the algorithm for preference application. The first stage of this investigation developed a characteristic feature vector to describe the physical appearance of an individual truss. The feature vector translates visual elements of a truss into quantifiable properties transparent to the computer algorithm. The nine elements in the feature vector were selected from an assortment of geometrical and behavioral factors and describe truss simplicity, general shape, and chord shape. Using individual feature vectors, a truss population may be divided into groups of similar design. Partitioning the population simplifies the feedback process by allowing users to identify groups that best suit their design preferences. Several unsupervised clustering mechanisms were evaluated for their ability to generate truss classifications that matched human judgment and minimized intra-group deviation. A one-dimensional Kohonen self-organizing map was selected. The characteristic feature vectors of truss designs within user-selected groups provided a basis for determining whether or not a user would like a new design. After analyzing user inputs, prediction algorithm trials sought to reproduce these inputs and apply them to the prediction of acceptable designs. This investigation developed a hybrid method combining rough set reduct techniques and a back-propagation neural network. This hybrid prediction mechanism was embedded into the operations of an Implicit Redundant Representation Genetic Algorithm. Locations within the ranking and selection processes of this algorithm formed the basis of a study to investigate the effect of user preference on truss optimization. Final results for this investigation prove that incorporating a user's aesthetic design preferences into the optimization project generates more design alternatives for the user to examine; that these alternatives are more in line with a user's conceptual perception of the project; and that these alternatives remain structurally optimal.

Bayesian optimization with Gaussian processes has become increasingly popular in the machine learning community. It is efficient and can be used when very little is known about the objective function, making it useful for optimizing expensive black box functions. We examine the case of using Bayesian optimization when the objective function requires feedback from a human. We call this class of problems \emph{interactive} Bayesian optimization. Here, we assume a parameterized model, and a user whose task is to find an acceptable set of parameters according to some perceptual value function that cannot easily be articulated. This requires special attention to the qualities that make this a unique problem, and so, we introduce three novel extensions: the application of Bayesian optimization to "preference galleries", where human feedback is in the form of preferences over a set of instances; a particle-filter method for learning the distribution of model hyperparameters over heterogeneous users and tasks; and a bandit-based method of using a portfolio of utility functions to select sample points. Using a variety of test functions, we validate our extensions empirically on both low- and high-dimensional objective functions. We also present graphics and animation applications that use interactive Bayesian optimization techniques to help artists find parameters on difficult problems. We show that even with minimal domain knowledge, an interface using interactive Bayesian optimization is much more efficient and effective than traditional "parameter twiddling" techniques on the same problem.

Currently, most of the big cities are facing the traffic congestion problem. Encouraging more people to use public transportation is considered as one of the best solutions (Hartley and Bargiela, 2001). To provide travellers with individual travel advice for journeys, today's Advanced Traffic Information Systems require efficient algorithms that are able to find the `best' route(s) for the public transport users according to their preferences. The aim of this research is to develop optimal and efficient time-dependent shortest path(s) algorithms which can accommodate public transportation users' preferences such as `minimum travel time', `minimum number of bus-changes', and `minimum walking time'. Two solutions are provided and compared in this report: the objective-oriented shortest path algorithms and the K-shortest paths (KSP) algorithm. Three different objective-oriented shortest path algorithms are used to accommodate three different user preferences separately. To find a route with `minimum travel time' (earliest arrival time or latest departure time), a forward/backward search algorithm is developed. The combination of the minimum transfer path search algorithm and the forward search algorithm is sufficient to find the path with the `minimum number of buschanges' (Huang and Peng, 2002). With a slight modification to the forwardibackward search algorithm, the algorithm can find a route that has `minimum walking time'. The K-shortest paths algorithm considers the three different user preferences simultaneously. It computes a reasonable number of ranked shortest paths, with the ultimate `most optimal' path being selected by consideration of the preferences. The experimental results based on the public transportation network of Nottingham City show that the objective-oriented shortest path algorithms work efficiently but can only optimise one single user preference. Using the KSP algorithm to accommodate user preferences is feasible. However, the execution time of the KSP algorithm is still long. A heuristic method is thus proposed to reduce the execution time by solving the `overlap ratio' between the k shortest routes. Experiment results are presented in Chapter Six

The urban woodlands in Redditch are integral elements of greenspace that are highly valued as an informal recreational resource, and for the physical and spiritual benefits that the community derives from visual and physical contact with them, on a daily basis. However, that they are cherished community assets, rather than places to be avoided, depends on contextually specific requirements being met, in regard to their siting and design. The thesis discusses these key criteria from the perspective of the Redditch urban community. It is identified that a choice of woods should be located within 300 and 700 metres of the home, to enable access by people, including those with restricted home range, mobility or with limited time to visit, and to provide a moderate walk to woods more distant, as an integral part of the 'recreational experience'. Convenient access to, and familiarity of these urban woodlands increases people's confidence to use woods more frequently and more distant from the home. Woodland size, preferably between 2 to 7 hectares and a good network of well lit, hard surfaced paths are also important factors influencing the attractiveness of woodland. Otherwise, visitors' demands regarding woodland type and facilities are modest. Plantations are as much valued and enjoyed as ancient semi-natural woodland. However even where the physical requirements for woodland are met, social and cultural factors limit many people's access to, and uses of the urban woodland. By exploring the personal, social and cultural values, and interpretations of these woods, the thesis analyses how the community reacts to change to the woodlands introduced by woodland management works, and identifies that it ascribes them a plurality of meanings and contextual relationships; as a woodland garden, a doorstep recreational area, a symbol of the pastoral idyll, a wildlife sanctuary and a gateway to the natural world. The theoretical framework of the thesis draws on multi-disciplinary perspectives including; landscape deslqn, town planning and the social and cultural perspectives of cultural geography. The evolution of the urban woodlands as elements of urban greenspace, people's recreational uses, and their attitudes and feelings towards them are explored by diverse methodological procedures, which include a longitudinal study and use of both qualitative and quantitative data collection methods. The research both builds on and adds to the existing body of knowledge by addressing the value and significance of providing urban woodlands within the urban fabric, and the key criteria which need to be observed to provide such areas close to where people live, and close to what they need.

Location-sharing services have become increasingly popular with the proliferation of smartphones and online social networks. People share their locations with each other to record their daily lives or satisfy their social needs. At the same time, inappropriate disclosure of location information poses threats to people's privacy. One of the reasons why people fail to protect their location privacy is the difficulty of using the current mechanisms to manually configure location-privacy settings. Since people's location-privacy preferences are context-aware, manual configuration is cumbersome. People's incapability and unwillingness to do so lead to unexpected location disclosures that violate their location privacy. In this thesis, we investigate the feasibility of using recommender systems to help people protect their location privacy. We examine the performance of location-privacy recommender systems and compare it with the state-of-the-art. We also conduct online user studies to understand people's acceptance of such recommender systems and their concerns. We revise our design of the systems according to the results of the user studies. We find that user-based collaborative filtering can accurately recommend location-privacy preferences and outperform the state-of-the-art when training data are insufficient. From users' perspective, their acceptance of location-privacy recommender systems is affected by the openness and the context of recommendations and their privacy concerns about the systems. It is feasible to use data obfuscation or decentralisation to alleviate people's concerns and meanwhile keep the systems robust against malicious data attacks.

This thesis describes a study carried out with the aim of discovering user preferences as to the design of 3-dimensional virtual worlds for accessing information. No literature was found which dealt with this topic, and it was therefore thought that, rather than ask users to make a selection from arbitrarily-chosen designs, it would be informative to consult the users from the beginning of the design process. To this end, a Grounded Theory methodology was adopted, and users were selected from postgraduate students and staff from Information Management courses at the Robert Gordon University, Aberdeen. Three “rounds” of interviews were conducted. The first round was concerned with finding out what ideas for a world design people would have, the second with testing four worlds derived from the first round, and the third with exploring further ideas that users had, based on their experience of the test worlds. At each stage of the process, emergent theories were constructed, and modified according to subsequent findings. It was established that the factors which influenced this group of users in their preferences for the design of worlds were not structural, as might have been assumed, but instead were related to properties such as familiarity, organisation, assistance, and quality of information and presentation. When the results were examined in the context of developments in the use of virtual environments, it was found that they provide a theoretical underpinning for practices such as the provision of “conventional” library structures in the popular online environment Second Life. This is not a statistical exercise, but it would appear that there are no significant differences based on the criteria of age, gender, or whether a user was staff or student. More thorough studies would be required to determine this absolutely, but for the moment it appears more useful to draw a broad set of conclusions. ii Issues were identified which indicate potentially rewarding areas for further research and design. Specifically, it would be of interest to discover whether the affective responses of these groups are also common to other groups, and to experiment further with worlds designed in the light of the current findings. Further investigation of the small number of cases in which users do not respond to the worlds would also be desirable, to determine whether this response is characteristic of a group of people who will not react positively to any world, or whether these users simply reacted negatively to the examples presented.

Playing music is considered essential for some businesses. When entering a clothing store, a café or a gym, there is most often some music playing in the background. The employees do not have the ability to select music optimally to maximize profit. Their expertise lies within their main duties of the workplace and they should spend most of their time focusing on those duties for an efficient workflow. The problem that arises is how businesses can play suitable music with minimal effort in music selection. To solve this, a recommender system is built with the real-time machine learning algorithm, DR-TRON. It is a lightweight and dynamic algorithm that instantly improves on user interaction. By using the dynamic nature of the algorithm, a more trivial model was initially built to test for some valuable output. Afterward, a more complex model was built where there was more consideration in music channel properties. The second model recommends suitable music channels and reduces the effort of selection.

Music has many benefits for our mood and feelings, especially so when we get to choose our own favorite music. However, accessing one's favorite music is not as easy for everyone. For motorically disabled and locked-in people, interacting with devices used for listening to music is challenging since it requires physical interaction. Machine learning classification methods used with EEG could prove useful for detecting individual musical preferences, extracted without any physical or verbal interaction. The two most common methods within EEG-based classification are Artificial neural networks (ANN) and Support vector machines (SVM). This study compares the performance of these two methods, when used with the DEAP dataset of EEG-monitored participants watching music videos. This comparison can help with gaining insight into which machine learning method is most appropriate for music preference detection, contributing towards more accurate predictions of motorically disabled people’s musical preferences. The participants in the DEAP dataset rated watched music videos by preference from 1 to 9, of which we trained the classification models to separate between higher (chosen as ratings 8 to 9) and lower ratings. From the results it is concluded that ANN performs better than SVM in terms of accuracy, with ANN performing at roughly 86% and SVM at 85%, while the SVMs were substantially faster to train. These accuracy scores were obtained from two ANN and SVM models using the optimal parameter and channel configurations, which were calculated through extensive testing. The accuracies are however likely achieved due to an imbalanced dataset, with too few data samples of higher ratings in proportion to lower, leading to biased classifiers that work well on our dataset but has probably close to random classification performance. Alterations to our methods could give better performing classifiers, and would also lead to more meaningful comparisons of ANN and SVM for EEG-based musical preference prediction.<br>Musik har många fördelar för vårt humör och våra känslor, synnerligen när vi får lyssna på vår favoritmusik. Det är däremot inte lika enkelt för alla att komma åt sin egen favoritmusik. För rörelseförhindrade och personer med inlåsningssyndrom är det utmanande att interagera med de enheter som används för att lyssna på musik eftersom att de kräver fysisk interaktion. Maskininlärningsmetoder för klassificering av EEG-data skulle kunna vara användbara för att upptäcka individuella preferenser av musik utan fysisk eller verbal interaktion. De två vanligaste metoderna inom EEG-baserad klassificering är Artificiella neurala nätverk (ANN) samt Stödvektormaskiner (SVM). Studien jämför prestanda av dessa metoder på DEAP-datasetet av EEG-övervakade deltagare, för att få insikt i vilken maskininlärningsmetod som är mest användbar för klassificering av musikpreferenser. Jämförelsen kan bidra till insikter om vilken maskininlärningsmetod som passar bäst för klassificering av musikpreferenser, vilket skulle kunna bidra till precisare klassificeringar av musikpreferenser bland personer med rörelsehinder. Deltagarna i DEAP-datasetet betygsatte musikvideorna utifrån preferens på en skala mellan 1 till 9, vilket användes för att träna klassificeringsmodellerna för att separera mellan högre (valda som betyg 8 till 9) och lägre betygsättningar. Från resultaten är det slutställt att ANN presterar bättre än SVM vad gäller noggrannhet, där ANN presterar runt 86% och SVM kring 85%, medan SVM var avsevärt snabbare att träna. Dessa noggrannheter erhölls från en ANN och SVM genom att använda de optimala parameter- och kanalkonfigurationer, vilka beräknades genom omfattande tester. Noggrannheterna är däremot troligtvis uppnådda på grund av ett obalanserat dataset, med för få datapunkter med högre betyg i proportion mot de lägre, vilket leder till partiska klassificerare som fungerar väl på vårt dataset men som troligtvis har en klassificeringsprestanda närmre slumpen. Ändringar i våra metoder skulle kunna ge bättre presterande klassificerare, och skulle också kunna leda till mer meningsfulla jämförelser av ANN och SVM för EEG-baserad klassificering av musikpreferenser.

Recommender systems assist users in finding what they want. The challenging issue is how to efficiently acquire user preferences or user information needs for building personalized recommender systems. This research explores the acquisition of user preferences using data taxonomy information to enhance personalized recommendations for alleviating cold-start problem. A concept hierarchy model is proposed, which provides a two-dimensional hierarchy for acquiring user preferences. The language model is also extended for the proposed hierarchy in order to generate an effective recommender algorithm. Both Amazon.com book and music datasets are used to evaluate the proposed approach, and the experimental results show that the proposed approach is promising.

碩士<br>明道大學<br>設計學院碩士班<br>103<br>Aromatherapy can be used to enhance the natural aroma of home atmosphere aromatherapy beauty therapy research, medical, emotional therapy, been recognized and confirmed its effectiveness many related units. Aromatherapy aids occupy an important role in modern life, with the life quality improvement, aromatherapy aids mushroomed released, in addition to pay attention to home aromatherapy, but also when they go out in a fragrant, this study, the most common aromatherapy fogging machine (water, oxygen machine), washing machine, dryer Diffuser, Diffuser bamboo, oil lamp, Diffuser stone, aromatherapy jewelry and other seven auxiliary equipment, study the pros and cons of using a difference, by 10 interviews with 150 people made on the basis of a questionnaire, is the focus of this study. The study found the following analytical data, assist with the future development of aromatherapy can have a better basis and direction, to create a more convenient and in line with public demand for alternative aromatherapy in the realm. First, seven incense assist with public believe assisting device modeling appearance is the first prerequisite. Second, the user would like to have the same good before using aromatherapy aroma release assisting device. Third, the hope aromatherapy aids low failure and can use more extended period. Fourth, to enhance the aroma of incense assisting device can both carry the decorations. Five, wanted to create a more ergonomic assist with the incense. Sixth, atomizing machine can be combined with the dual function of washing machine aromatherapy aids. More of the research data can be used as back aromatherapy aids the design basis and the reference direction.

Many services fail. Failures are those encounters during which the user assesses the service as flawed (Maxham & Netemeyer, 2003; Palmer, Beggs, & Keown-McMullan, 2000), or improper (Maxham, 2001). An emergent service failure literature has determined that such failures play an important role in user perceptions and subsequent behaviours. The present study sought to discover the ways in which possible users responded to an ambiguous service failure. The purpose of the study was to twofold: (a) to understand the situational, emotional, and dispositional factors which influence users??? likelihood of complaining behaviours; (b) to develop connections between these factors and users??? preferences for service recovery. We were guided by several research questions including: RQ1: When the cause of the failure is ambiguous, to what/whom do users attribute blame? RQ2: How well do attribution of blame and perceived failure severity predict negative emotions? RQ3: What is the relative influence of each negative emotion in predicting the likelihood of each complaining behavior? RQ4: Does user disposition intervene in the relationship between emotion and complaining? RQ5: What is the role of appraisals and emotion in determining service recovery preferences? Borrowing from the service quality, consumer behaviour, health, and therapy literatures, we develop a conceptual framework for answering our questions. Consistent with recent research, we conceptualized that negative emotions following service failures were dependent on users??? appraisals of the situation. Specifically, we looked at the effects of failure severity and attribution of blame on anger, frustration, shame, guilt, regret, and dissatisfaction. We then conceptualized complaining behaviour as a coping mechanism for these negative emotions, and preferences for service recovery as manifestations of immediate desires users would have to address their emotions. To test our theory, an experiment with hypothetical scenarios and a survey instrument was developed. We manipulated two conditions (time lost and money lost) at different points in the survey while participants while self-selected into an attribution of blame condition (Self, provider, other). Undergraduate students at the University of Waterloo (n=288) served as the sample. The questionnaire assessed such variables as attitudes towards complaining, locus of control, tendency for avoidance, emotional response, complaining behaviours, and preferences for service recovery. Results from multivariate analyses confirmed that appraisals help predict negative emotions, and that negative emotions influence complaining behaviours. Results also demonstrated that appraisals and emotions do begin to explain variance in service recovery preferences. Contrary to the interactionist approach, results failed to support the notion that personal dispositions (such as attitudes and personality traits) moderate the relationship between situational factors and behaviour. Finally, conclusions for the study are made, and implications for future research and the design of service recovery strategies are discussed.

碩士<br>中華大學<br>資訊工程學系碩士班<br>87<br>Abstract With the explosive growth of on-line information on various kinds of platforms (e.g. the library and the Internet), users may have more opportunities to get information conveniently. However, without an efficient and personalized information retrieval (IR) system, the users could also be misleader in the huge amount of information. Therefore, how to search for useful information for each individual users has become an urgent problem for most researchers in IR, library science and the Internet. To attack the problem, IR systems should be able to consider different users' information needs. In this thesis, I propose a model for building adaptive information retrieval (AIR) systems which may adapt its search strategies to users' different information needs and preferences. Thus, both the quality and the efficiency of IR may be promoted. In particular, AIR should be able to (1) recognize the personal information needs of individual users, (2) detect the change of user preferences, and (3) map the user’s semantic preferences to the contents of information. However, according to the survey conducted in the thesis, all search engines on the Internet are unable to achieve the three tasks. Because users' information needs and preferences are often expressed in their queries, AIR should observe and acquire users' semantic preferences from their queries. The output of the learning module is the mapping between query terms and their suitable subset of the document database. As next query is entered, its terms are extracted and mapped to suitable document database in which useful information is more likely to be found. AIR employs a set of the Heuristic Semantic Patterns (HSPs) to learn users' semantic preferences. The HSPs works on a tree-structured document database which is common for most libraries and web sites on the Internet. A semantic preference of a term is expressed as a mapping between the term and its suitable document category on the document database. As more semantic preferences of query terms may be acquired, AIR may adapt its search strategy to individual needs and preferences. In this thesis, AIR is explored both theoretically and empirically. The impacts and contributions of the work will be evaluated in terms of the extent to which both the quality and the efficiency of the IR are improved.