Dissertations / Theses on the topic 'Keyword-based search'

Thesis (Ph. D.) -- University of Maryland, College Park, 2006.
Thesis research directed by: Computer Science. Title from t.p. of PDF. Includes bibliographical references. Published by UMI Dissertation Services, Ann Arbor, Mich. Also available in paper.

The work conducted in this thesis aims to present an hybrid model for a real­ world application search engine. The project presented was part of an intern­ship work carried out in a start­up which deals with Knowledge Management and Artificial Intelligence. The aim of the internship work was to improve the current search engine system to build a new system for a future web ap­plication use case. An in­-depth study on the limitations of keyword search alone, and on semantic search, revealed the need of a transition from a pure keyword­-based information retrieval system to an hybrid model, making use of both keyword search and semantic search. In particular the old system re­lied on a tfidf­-based algorithm, while the final model tries to overcome the limits of keyword search by joining the abilities of OkapiBM25, a probabilis­tic information retrieval approach, with newer semantic search models based on SentenceBERT. The models, and the algorithm implemented, exploit deeply recent techniques in Information Retrieval such as lexical search, sim­ilarity search, query expansion, document expansion and automatic question generation. The data used to test the models came from a banking dataset, be­longing to one of the company clients, previously created for an Information Retrieval chat­bot. Different experiments led to a final model able to improve the search performances showing great advantages with respect to keyword search and pure semantic search.

In this paper, we present a P2P-based database sharing system that provides information sharing capabilities through keyword-based search techniques. Our system requires neither a global schema nor schema mappings between different databases, and our keyword-based search algorithms are robust in the presence of frequent changes in the content and membership of peers. To facilitate data integration, we introduce keyword join operator to combine partial answers containing different keywords into complete answers. We also present an efficient algorithm that optimize the keyword join operations for partial answer integration. Our experimental study on both real and synthetic datasets demonstrates the effectiveness of our algorithms, and the efficiency of the proposed query processing strategies.
Singapore-MIT Alliance (SMA)

Technology keeps on evolving and so must the science of information retrieval. This thesis presents keyword-based approaches to improve information retrieval from the Internet. Focused and unfocused queries to search engines are considered, and means of obtaining relevant documents are presented. For focused queries, techniques are provided to obtain a high precision score from the hit documents; these documents do contain the exact answers to the focused query, which is usually a question. User queries are subjected to ambiguity test to determine if it is ambiguous, and if it is so, provide direction so as the user's intended meaning is the one that is actually searched. The queries are modified to form a new clear and unambiguous. Query is sent to several search engines at the same time, and hit documents from each of these search engines are collated and merged. Hit documents to an ambiguous query are analyzed and ranked based on their actual relevance to the query. Term frequency is used, along with popularity score, to determine the total score of a relevant document. Every relevant hit document is classified based on its academic relevance. A few academic categories are considered--(1) Course Notes, (2) Frequently Asked Questions, (3) Research Paper, (4) Technical Report, (5) Thesis, (6) Tutorial, (7) Review, and (8) Research Paper/Technical Report. Once a search is done, a set of relevant documents is presented, along with each document's academic relevance category (if any)

碩士
國立臺灣大學
資訊管理研究所
92
Abstract Peer-to-peer search is to search for objects on a peer-to-peer network according to application-defined semantics. It is a challenging problem because the advantages of a peer-to-peer network are to be preserved while feasible performance must be achieved. A DHT is a distributed name resolution architecture for a peer-to-peer network of Internet-scale; however, it does not possess search ability. Some work tries to build global inverted indexes on a DHT in order to provide keyword search ability, but these approaches suffer from severe unbalanced load at each node. It is due to the power-law nature of any document set. Semantic routing, in which each node based on locally collected information forwards received queries to only those neighbors that are semantically close to the queries, is another category of solutions to peer-to-peer search problems. These approaches often adopt sophisticated information retrieval algorithms to guide message forwarding on Gnutella-like peer-to-peer networks. Semantic routing is more precise in terms of document retrieval than global inverted indexing, but is inefficient when the network size scales to the Internet size. We propose a keyword search scheme on DHT-based peer-to-peer network in order to scale our solution to the Internet size. Our scheme is to index objects according their associated keyword sets with Bloom filter techniques. Each object is represented as an r-bit filter vector. In the meantime, we build an r-dimensional hypercubic keyword search layer on top of a DHT. Then we place each object on the hypercube according to its filter vector. Our approach is proved to provide the function of a global inverted index but achieves much better load balance than previous work even under the power-law nature of an object set. The hypercubic keyword search layer also implies index closeness in its topology, which improves search performance when only a portion of query-satisfying objects are needed. Each query made can be resolved within logarithmic hop counts, and the cost to resolve a query is linear to the resultant set size of the query.

碩士
國立臺灣科技大學
資訊工程系
105
Cloud storages have gained popularity in the recent years. With the increasing quantity of data outsourced to cloud storages, keyword search over encrypted cloud data with the consideration of privacy preservation has become an important topic. The majority techniques in the literature only provide exact single or multiple keyword search in which the keywords have to exactly match those in a pre-defined dictionary. However, restricting users’keywords within the pre-defined dictionary is impractical for real-world applications. Some existing fuzzy keyword search schemes only focus on dealing with spelling mistakes of keywords. The flexibility of keywords used in the search is not considered. This paper addresses the problem of semantic multi-keyword search over encrypted cloud data. Users can use keywords not just in the pre-defined dictionary of the dataset, but with the flexibility of their own choice. The similarity of the given keywords with the search index of each document is then calculated. An adequate set of documents are selected as the results for the search based on the similarity. In addition, privacy of the search is also considered during the search executed by the third party service provider. Experiments are conducted using a dataset of massive papers in real world. The experimental analyses show that the proposed scheme can perform the semantic multi-keyword search effectively and efficiently.

碩士
中原大學
資訊工程研究所
96
XML is becoming the standard for data exchanges and data representation. Keyword search allows a user to easily query and access XML data and thus becomes an important research issue. In this thesis, we used Dewey-base two indices to index XML data. It does not require re-computation of existing labels when an XML document is updated. When returning keyword search results, we use the relationships between nodes ensure the return data is correct and to avoid returning meaningless data. Users can query XML data without having to understand the XML structure. An example system was implemented as an illustration.

碩士
清雲科技大學
資訊管理所
100
According to Nielsen Online research report, children online population grew 18 percent over the past five years. Storybooks have been the main source of children reading. With the advance of digital contents and computer technologies, how young children choose their interest digital storybooks in a digital library or on the Internet is becoming important. However, children often encounter difficulties in finding books with traditional keyword search interfaces due to the immaturity of cognitive development. Therefore, the purpose of this study is to develop an intuitive, iconic keyword search interface with which children compose comprehensible image keywords for their intended storybooks. This study also investigates the usability of the iconic keyword-based storybook search system for young children. The evaluation of the system involved 20 kindergarten and 20 second grade elementary children. The evaluation results showed that kindergarten children composed their search terms mainly based on characters on the book covers in both tasks of finding favorite storybooks and searching specific storybooks. However, the elementary children showing their imagination and observation demonstrated more detailed search terms by including elements of scenes of a book cover in the aforementioned tasks. For the usability aspect, elementary children showed higher rates of task completion and correctness of usability questions.

碩士
國立中山大學
資訊工程學系研究所
107
With the rapid advancement of cloud computing, users upload their files to the cloud server so that any user can access it remotely. Naturally, the security and privacy concerns take place in such scenario since we store files in a remote server which is not fully trusted. Therefore, assurance of the security and privacy of remote data becomes an important aspects to the researchers. To assure the data safety, the data owner, typically, encrypts the data before outsourcing them to the cloud server. In addition, an encryption mechanism needs to enable the consumers to perform efficient searches of such encrypted data in the cloud storages through keywords. This is commonly known as searchable encryption. There is a lot of attention to design a reliable searchable encryption in recent days, however, most of them are improper due to several limitations, such as the requirement of an on-line fully trusted third party, poor efficiency, high-overhead in user revocation, support of a single keyword search, etc. To mitigate such limitations, an attribute-based encryption scheme with fine-grained multi-keyword search is proposed. The new scheme supports the user revocation. In addition, the length of the ciphertext as well as the secret key do not grow linearly under the influence of the size of attribute set. The performance of the proposed scheme is better as compared to other related schemes. Hence, one can easily adopt the proposed scheme for the real life applications due to its flexibility in terms of its features, security and efficiency.

碩士
長庚大學
資訊管理研究所
95
Peer-to-peer systems are a special type of decentralized systems which don’t need centralized control and class hierarchies. Current peer-to-peer applications have the following characteristics: storage, efficiency, anonymity, searches, credibility and the advantages of the class naming. P2P networks can be divided into two types: structured and unstructured. The structured P2P networks ,using a hash function to map a resource to a node IP address where the resource is to be placed. Besides, rapid routing mechanisms are designed to facilitate resource lookups. However, because structured P2P networks exploit hash functions to share and search resource, the cryptographic characteristics of hash functions render keyword-based searches unfeasible in structured P2P networks. This research aims to provide a novel search mechanism for structured Peer-to-Peer networks to search by keywords and ranges. Our simulation results show the proposed mechanism enables a more efficient keyword-based search in the system which use hash functions share and search resource directly , as well as maintaining a higher level of availability in the mechanism which provide search by ranges.

碩士
國立中興大學
資訊科學與工程學系
103
The UCIS-X (An Updatable Compressing and Indexing Scheme for XML) indexing method uses the Dewey encoding to record the parent-child relationship of nodes in an XML file, and the branch information between XML nodes are encoded in Branch maps. UCIS-X outperforms other XML indexing schemes in terms of index space and query response time. But its performance will be downgraded in case of keyword search in XML file with large contents due to linear list used the content-index. In this thesis, we modified the Content-Index of UCIS-X by designing a hashed content-index using content keywords as hash keys to improve the performance on keyword search. The experimental results show that the proposed method has about 44% performance improvement over UCIS-X on the content-rich benchmark dataset XMark, and it also has about 65.5% performance improvement over UCIS-X on structure-duplicated dataset DBLP.

碩士
國立交通大學
資訊管理研究所
93
Peer-to-Peer (P2P) network architecture proposes a new technology to exchange information and service between computers without centralized controllers. These resources and service including : files sharing, instant messaging, and distributed computing . Now, P2P file sharing is one of the P2P hottest applications. This paper proposes a P2P network based on semantic keyword search. It is easy to implement semantic keyword search in “Centralized Network Architecture” because we can put the semantic knowledge base on the centralized server ; however, there is no peer always in the system. Therefore, our main concern is to integrate the advantage of “Super Peer Network” and the “Query Expansion”(an information retrieval method in centralized server-to-client architecture) .Our approach can solve the problems mentioned before. According to the simulation, our approach can provide good performance of semantic keyword searching in P2P system and solve KRDB synchronization efficacious.

碩士
國立雲林科技大學
資訊管理系
103
Keywords are a subset of words or phrases from a document those can describe the meaning of the document. The major methods for Chinese keyword extraction are keyword lexicons approaches, statistics approaches, linguistics approaches, etc. Among these methods, keyword lexicons approaches make keyword extraction high precision and high efficient, but building keyword lexicons spends a lot of time and the maintenance of keyword lexicons is manual. This research presents a Chinese keyword extraction system based on CKIP Chinese word segmentation system. This system provides the recombination of words by using part of speech (POS) combination and automatic words combination via search engine (Google Search) and internet encyclopedia (Wikipedia). This system also focuses on building a keyword lexicon that can update its keywords automatically. The system can improve the disadvantages of keyword lexicons approaches. The results of experiments show that using the CKIP Chinese word segmentation system, POS combination and automatic words combination gains higher precision and the number of documents does not affect the performance of the keyword extraction system. Keywords: Keyword Extraction, Keyword Lexicon, Search Engine, Internet Encyclopedia

碩士
國立屏東科技大學
資訊管理系
90
In this epoch, the growth of information technology has exploded. How to mine the information of people’s interest in a short time has become an important issue. In general, people need to record addresses of every search engine before they begin information retrieval and adapt to different interfaces that every search engine provides. After getting each retrieve results, people need to filter the information that has filled with high repetition from independent search engines which set keywords or query conditions. This study develops an effective search mechanism to help people to find out the required information. It also aims to implement a Hyperlink and Keyword Frequency Analysis (HKFA) search engine. The HKFA search engine consists of five modules. There are (1)Query Module; (2)Information Retrieval Module─includes Dispatcher Algorithm; (3)Information Extraction Module─includes Feature Extraction Algorithm and Hyperlink Normal Form Algorithm; (4)Information Filtering Module─includes Occurrence Hit Algorithm、Filter Hyperlink Algorithm、Keyword Frequency and Position Algorithm; (5) Information Integration Module. The search engine mechanism designed in this study is to provide high precision and high recall information. Furthermore, this mechanism can not only effectively save people’s time spent on retrieving and filtering information but also reduce user’s information overhead.

碩士
國立嘉義大學
資訊工程學系研究所
107
Nowadays, cloud security is an important field and one of the topics in it is about secure cloud storage. When users send data to the cloud system, they cannot know the status of stored data and whether someone else can access their files or not. Therefore, the encryption of files stored in the cloud service provider is a very important requirement for data privacy. In this thesis, we apply the appropriate encryption method, attribute-based encryption, for solving the confidentiality problems in the cloud. Attribute-based encryption is based on attributes with highly flexible access control. The owner of the file establishes an attribute access structure to encrypt the file. The user possesses an attribute sets for himself. If user's attribute satisfies the ciphertext access structure, the user has the ability to decrypt. This method is very suitable for cloud storage. The keyword search based on the encrypted file is also an important issue. We integrate the attribute encryption and linear secret sharing scheme (LSSS) to generate an efficient keyword query token. The cloud service provider compares the keywords of the ciphertext and the keywords query token issued by the user without knowing the plaintexts of the file and the corresponding keywords. If the comparison result is valid, the cloud service provider will send the ciphertext to the user for decryption. Although applying LSSS to the attribute-based cryptosystem, the file owner needs to generate a matrix as encryption, it is relatively fast in decryption of the ciphertext. Therefore, the LSSS can be suitable for the cloud service provider as well as the various mobile devices. In the cloud system with attribute-based encryption, revoking attribute is a flexible approach to changing the access permissions. In our system, the re-encryption mechanism can properly be used to achieve the goal.

碩士
國立臺灣海洋大學
資訊工程學系
98
Users can search for the information they need by file name in Peer-to-Peer (P2P) networks. When the information we have is not sufficient to fully describe an object, then it is possible that the object will not be found. Therefore, keyword search, which is a flexible search service, has been developed. It is hashing the object’s keyword set to obtain a key that maps to a unique node in the network. File name search is hashing the object’s name to obtain a key. As a result, keyword search is more convenient and practical than file name search in finding an object. The hypercube structure has been proposed to be used in keyword search. Each object is mapped a vector according to it’s keyword set, and stored in the node corresponding to the vector. Objects with similar keywords are likely to be stored in close nodes, thus provides an efficient search scheme. When a search request simply demands some nodes in response, search message of routing strategy will determine that the message through which node, and hence affect the search efficiency, the amount of messages, and the node load balancing. We propose a random routing strategy based on hypercube structure in keyword search. It’s shown the request can be processed evenly through each node, and objects containing user’s keyword will have equal opportunity to be found.

Since the beginning of the 21st century, service-oriented architecture (SOA) has emerged as an advancement of distributed computing. SOA is a framework where software modules are developed using straightforward interfaces, and each module serves a specific array of functions. It delivers enterprise applications individually or integrated into a more significant composite Web services. However, SOA implementation faces several challenges, hindering its broader adaptation. This thesis aims to highlight three significant challenges in the implementation of SOA. The abundance of functionally similar Web services and the lack of integrity with non-functional features such as Quality of Service (QoS) leads to the difficulties in the prediction of QoS. Thus, the first challenge to be addressed is to find an efficient scheme for the prediction of QoS. The use of software source code metrics is a widely accepted alternative solution. Source code metrics are measured at a micro level and aggregated at the macro level to represent the software adequately. However, the effect of aggregation schemes on QoS prediction using source code metrics remains unexplored. The inequality distribution model, the Theil index, is proposed in this research to aggregate micro level source code metrics for three different datasets and compare the quality of QoS prediction. The experiment results show that the Theil index is a practical solution for effective QoS prediction. The second challenge is to search and compose suitable Web services with- out the need for expertise in composition tools. Currently, the existing approaches need system engineers with extensive knowledge of SOA techniques. A keyword-based search is a common approach for information retrieval which does not require an understanding of a query language or the underlying data structure. The proposed framework uses a schema-based keyword search over the relational database for an efficient Web service search and composition. Experiments are conducted with the WS-Dream data set to evaluate Web service search and composition framework using adequate performance parameters. The results of a quality constraints experiments show that the schema-based keyword search can achieve a better success rate than the existing approaches. Building an efficient data architecture for SOA applications is the third challenge as real-world SOA applications are required to process a vast quantity of data to produce a valuable service on demand. Contemporary SOA data processing systems such as the Enterprise Data Warehouse (EDW) lack scalability. A data lake, a productive data environment, is proposed to improve data ingestion for SOA systems. The data lake architecture stores both structured and unstructured data using the Hadoop Distributed File System (HDFS). Experiment results compare the data ingestion time of data lake and EDW. In the evaluation, the data lake-based architecture is implemented for personalized medication suggestion system. The data lake shows that it can generate patient clusters more concisely than the current EDW-based approaches. In summary, this research can effectively address three significant challenges for the broader adaptation of SOAs. The Theil index-based data aggregation model helps QoS prediction without the dependence on the Web service registry. Service engineers with less knowledge of SOA techniques can exploit a schema-based keyword search for a Web service search and composition. The data lake shows its potential to act as a data architecture for SOA applications.

2012 Joint 6th International Conference on Soft Computing and Intelligent Systems (SCIS) and 13th International Symposium on Advanced Intelligent Systems (ISIS) (SCIS-ISIS 2012). November 20-24, 2012, Kobe, Japan

While there are powerful keyword search systems that index all kinds of resources including emails and web pages, people have trouble recalling semantic facts such as the name, location, edit dates and keywords that uniquely identifies resources in their personal repositories. Reusing information exasperates this problem. A rarely used approach is to leverage episodic memory of file provenance. Provenance is traditionally defined as "the history of ownership of a valued object". In terms of documents, we consider not only the ownership, but also the operations performed on the document, especially those that related it to other people, events, or resources. This thesis investigates the potential advantages of using provenance data in desktop search, and consists of two manuscripts. First, a numerical analysis using field data from a longitudinal study shows that provenance information can effectively be used to identify files and resources in realistic repositories. We introduce the Leyline, the first provenance-based search system that supports dynamic relations between files and resources such as copy/paste, save as, file rename. The Leyline allows users to search by drawing search queries as graphs in a sketchpad. The Leyline overlays provenance information that may help users identify targets or explore information flow. A limited controlled experiment showed that this approach is feasible in terms of time and effort. Second, we explore the design of the Leyline, compare it to previous provenance-based desktop search systems, including their underlying assumptions and focus, search coverage and flexibility, and features and limitations.
Graduation date: 2012