Academic literature on the topic 'Multinomial mixture model'

Social media has become an integral part of the Internet. There have been users across the world sharing content like images, texts, videos, and so on. There is a huge amount of data being generated and it has become a challenge to the social media platforms to group the content for further usage like recommending a video. Especially, grouping videos based on similarity requires extracting features. This thesis investigates potential approaches to extract features that can help in determining the similarity between videos. Features of given videos are extracted using Object Detection and Action Recognition. Bag-of-features representation is used to build the vocabulary of all the features and transform data that can be useful in clustering videos. Probabilistic model-based clustering, Multinomial Mixture model is used to determine the underlying clusters within the data by maximizing the expected log-likelihood and estimating the parameters of data as well as probabilities of clusters. Analysis of clusters is done to understand the genre based on dominant actions and objects. Bayesian Information Criterion(BIC) and Akaike Information Criterion(AIC) are used to determine the optimal number of clusters within the given videos. AIC/BIC scores achieved minimum scores at 32 clusters which are chosen to be the optimal number of clusters. The data is labeled with the genres and Logistic regression is performed to check the cluster performance on test data and has achieved 96% accuracy

Ce travail de recherche traite de la problématique des transports dans les villes d’Indonésie, Jakarta et Bandung, mais également de la grande concurrence modale du trajet Jakarta-Bandung et Bandung-Jakarta. Les préférences des passagers sont des variables très importantes à connaître en raison de leurs impacts pour choisir un mode de transport parmi d’autres. Dans les transports, le modèle Logit est largement utilisé comme une méthode pour aborder la problématique du choix de transport multimodal comportant de multiples variables, mais dans la présente recherche, ces modèles ne sont pas appropriés pour la résolution de nos problèmes, car il y a des variables particulières à identifier et à prendre en compte. Par conséquent, nous avons développé pour nos besoins le modèle « Logit Mixed Multinomial Adapté (LMMA) » comme outil dédié à l’analyse décisionnelle dans le choix des modes de transport des passagers. La première partie de nos travaux de recherches porte sur l’identification et la compréhension des problèmes de transports intra-cité d’origine et de destination pour le choix du mode de transport entre Jakarta et Bandung (et puis entre Bandung et Jakarta). La seconde partie concerne le processus de décision final en proposant et en analysant les résultats d’un questionnaire adressé à de nombreux utilisateurs de la liaison Jakarta-Bandung (et Bandung-Jakarta). L’analyse permet pour chaque situation d’origine et de destination, et en tenant compte des services offerts par chaque mode inter-cité, d’identifier quel est le mode le plus compétitif<br>An ideal city or intercity transport system is one where all the transport networks, involving in general different modes of transport, could serve together the cities connections to fulfill a passenger demand and satisfaction. Each transport network should have a logical layout (as possible with minimum discontinuities) to meet the required demands. Also in that ideal system, the different modes of transport should not only have their own good performances but also the exchange between modes should be done with harmony. The conditions as mentioned above are worldwide challenges. The present work deals with the transportation problematic between two Indonesian cities, and also with the high modal competition on the Jakarta-Bandung corridor. On that corridor, road transport is currently the main demanding mode for passengers transportation. The airlines cannot compete and discontinued their operations to this route. Nowadays, railway transport is decaying. Passengers preferences are the main variables for the final modal choice. It is necessary to know preferences due to their decisions impacts to choose one mode over the others. Those preferences are in fact not simple to express in a complex city and intercity transport system. In transportation, the Logit model is widely used as a method to explore the problematic of modal choices involving a lot of different variables. There are several Logit models already developed, such as “General Extreme Value”, “Probit”, and “Nested model”, but in this research, they are not compatible to solve our defined problems because there are some particular identified variables to be taken into account. Therefore we propose the "Adapted Mixed Multinomial Logit (AMML)" Model as a tool for analysis towards passenger's decision in modal choices. On the Jakarta-Bandung corridor, modal choices are influenced by the encountered problems in intercity transport at origin and destination. One part on this research deals with identification and understanding of the intracity transport problems of origin and destination on the choice of transport mode in Jakarta-Bandung corridor (Jakarta-Bandung and Bandung-Jakarta direction). The second part of this research deals with the final decision process by analyzing the results of questionnaires addressed to many users of the Jakarta-Bandung corridor. The five main variables of the last questionnaire are travel time, overall cost, security conditions, quality of travel information and connectivity conditions relevant to intercity transport and intracities transport conditions as well. After validation of the questionaires, this research uses the AMML model to get final decision result by comparing one mode among three intercity transport mode (train, minibus, and car) using the values of the variables. Taking into account the characteristics of each intercity mode of transportation, the analysis identifies the most competitive intercity transport mode for each situation from departure city to arrival city. Using alternative public and private transport modes policies, one could in the future modify passenger choice on intercity transport mode. Therefore, this study is relevant for improving of intracity and intercity transport systems

A investigação em analise de agrupamento (cluster analysis) continua em curso. Identificar o número de grupos, bem como seleccionar um subconjunto de variáveis relevantes a partir de dados de uma amostra constituem domínios de investigação ativa em agrupamento. Grande parte dos métodos desenvolvidos para abordar estas temáticas refere-se a dados contínuos, e não podem ser directamente aplicados ao agrupamento de dados categoriais. Este trabalho, pretende ser um contributo nesta área, abordando o agrupamento de dados categoriais.<br>Research on cluster analysis continues to develop. Identifying the number of clusters and selecting a subset of relevant variables available in the data have been active areas in research on clustering methods. The approaches proposed for addressing these issues are mostly designed to deal with numerical data and cannot be directly applied for clustering categorical data. This work intends to be a contribution to handling categorical data, in this area.

Thesis (M.Sc.)--York University, 2008. Graduate Programme in Computer Science.<br>Typescript. Includes bibliographical references (leaves 88-90). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&res_dat=xri:pqdiss&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&rft_dat=xri:pqdiss:MR51574

<p>This thesis develops Bayesian latent class models for nested categorical data, e.g., people nested in households. The applications focus on generating synthetic microdata for public release and imputing missing data for household surveys, such as the 2010 U.S. Decennial Census.</p><p>The first contribution is methods for evaluating disclosure risks in fully synthetic categorical data. I quantify disclosure risks by computing Bayesian posterior probabilities that intruders can learn confidential values given the released data and assumptions about their prior knowledge. I demonstrate the methodology on a subset of data from the American Community Survey (ACS). The methods can be adapted to synthesizers for nested data, as demonstrated in later chapters of the thesis.</p><p>The second contribution is a novel two-level latent class model for nested categorical data. Here, I assume that all configurations of groups and units are theoretically possible. I use a nested Dirichlet Process prior distribution for the class membership probabilities. The nested structure facilitates simultaneous modeling of variables at both group and unit levels. I illustrate the modeling by generating synthetic data and imputing missing data for a subset of data from the 2012 ACS household data. I show that the model can capture within group relationships more effectively than standard one-level latent class models.</p><p>The third contribution is a version of the nested latent class model adapted for theoretically impossible combinations, e.g. a household with two household heads or a child older than her biological father. This version assigns zero probability to those impossible groups and units. I present a proof that the Markov Chain Monte Carlo (MCMC) sampling strategy estimates the desired target distribution. I illustrate this model by generating synthetic data and imputing missing data for a subset of data from the 2011 ACS household data. The results indicate that this version can estimate the joint distribution more effectively than the previous version.</p><br>Dissertation

JEL: C25, C52, M31<br>Actualmente, os modelos de Mistura são considerados um dos métodos de segmentação mais eficientes na área de Marketing para o estudo das estruturas de preferências. Com base numa amostra de clientes de uma Instituição Financeira, numa primeira fase, será realizada uma segmentação com base nos modelos de mistura finita, de forma a perceber as estruturas de necessidades de produtos financeiros. Com base nas perfilagens dos segmentos, será possível efectuar a avaliação da necessidade ou não de desenvolvimento de estratégias diferenciadas consoante os segmentos obtidos de forma a aumentar o valor da rendibilidade dos clientes já existentes, adequando assim a oferta de produtos. Nesta fase, serão obtidas as probabilidades de pertença a posteriori para classificar novos clientes nos segmentos mais adequados, permitindo contactar o cliente da melhor forma e com a melhor oferta. Numa segunda fase, será utilizado o modelo de mistura de regressões para perceber o impacto das acções de Marketing nos produtos detidos pelos clientes. Admitindo a existência de heterogeneidade das necessidades financeiras dos clientes, e colocando a hipótese de que as mesmas são explicadas com base nas acções de marketing realizadas, pretende obter-se um conjunto de estimativas de regressão para cada segmento identificado. A obtenção dessas estimativas de regressão, consoante a significância estatística, irá fornecer um maior conhecimento sobre a adequação da actual estratégia de marketing definida, e perceber a necessidade de afinação ou não da mesma consoante o segmento.<br>Nowadays, finite mixture models are one of the most efficient segmentation technique in the marketing field, in order to analyse structures of preferences in a given population. Based on a sample of clients of a given Financial Institution, the first step of this study applies a finite mixture model to understand the existing structures of financial needs of the clients. Based on the profiled segments, the need of developing different marketing strategies for each segment will be assessed, in order to increase the profit of the actual clients due to a correct contact strategy and offer of products. The probabilities of belonging to a certain segment will be obtained in order to alocate new clients in the most adequate segment, allowing to reach the clients with the best contact and offer of products strategy. In a second step, a regression mixture model will be applied to understand the impact of the actual marketing strategy in the portfolio of products of the clients. Assuming the existence of the heterogeneity in the financial needs of the clients and the fact that these needs can be explained by the acquisition campaigns, a set of regression models are estimated for each segment. Depending on the significance of this regression estimates, one understands the adequacy of the actual defined marketing strategy and decides if there is the need of improvement depending on the segment.