Search (654 results, page 33 of 33)

3.4805894E-4 = product of:
  0.005220884 = sum of:
    0.005220884 = product of:
      0.010441768 = sum of:
        0.010441768 = weight(_text_:information in 1154) [ClassicSimilarity], result of:
          0.010441768 = score(doc=1154,freq=14.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.20526241 = fieldWeight in 1154, product of:
              3.7416575 = tf(freq=14.0), with freq of:
                14.0 = termFreq=14.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=1154)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

In this thesis, we will present methods for introducing ontologies in information retrieval. The main hypothesis is that the inclusion of conceptual knowledge such as ontologies in the information retrieval process can contribute to the solution of major problems currently found in information retrieval. This utilization of ontologies has a number of challenges. Our focus is on the use of similarity measures derived from the knowledge about relations between concepts in ontologies, the recognition of semantic information in texts and the mapping of this knowledge into the ontologies in use, as well as how to fuse together the ideas of ontological similarity and ontological indexing into a realistic information retrieval scenario. To achieve the recognition of semantic knowledge in a text, shallow natural language processing is used during indexing that reveals knowledge to the level of noun phrases. Furthermore, we briefly cover the identification of semantic relations inside and between noun phrases, as well as discuss which kind of problems are caused by an increase in compoundness with respect to the structure of concepts in the evaluation of queries. Measuring similarity between concepts based on distances in the structure of the ontology is discussed. In addition, a shared nodes measure is introduced and, based on a set of intuitive similarity properties, compared to a number of different measures. In this comparison the shared nodes measure appears to be superior, though more computationally complex. Some of the major problems of shared nodes which relate to the way relations differ with respect to the degree they bring the concepts they connect closer are discussed. A generalized measure called weighted shared nodes is introduced to deal with these problems. Finally, the utilization of concept similarity in query evaluation is discussed. A semantic expansion approach that incorporates concept similarity is introduced and a generalized fuzzy set retrieval model that applies expansion during query evaluation is presented. While not commonly used in present information retrieval systems, it appears that the fuzzy set model comprises the flexibility needed when generalizing to an ontology-based retrieval model and, with the introduction of a hierarchical fuzzy aggregation principle, compound concepts can be handled in a straightforward and natural manner.

3.4805894E-4 = product of:
  0.005220884 = sum of:
    0.005220884 = product of:
      0.010441768 = sum of:
        0.010441768 = weight(_text_:information in 2567) [ClassicSimilarity], result of:
          0.010441768 = score(doc=2567,freq=14.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.20526241 = fieldWeight in 2567, product of:
              3.7416575 = tf(freq=14.0), with freq of:
                14.0 = termFreq=14.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=2567)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

As the Web continues to grow and encompass broader and more diverse sources of information, providing effective search facilities to users becomes an increasingly challenging problem. To help users deal with the deluge of Web-accessible information, we propose a search system which makes use of context to improve search results in a scalable way. By context, we mean any sources of information, in addition to any search query, that provide clues about the user's true information need. For instance, a user's bookmarks and search history can be considered a part of the search context. We consider two types of context-based search. The first type of functionality we consider is "similarity search." In this case, as the user is browsing Web pages, URLs for pages similar to the current page are retrieved and displayed in a side panel. No query is explicitly issued; context alone (i.e., the page currently being viewed) is used to provide the user with useful related information. The second type of functionality involves taking search context into account when ranking results to standard search queries. Web search differs from traditional information retrieval tasks in several major ways, making effective context-sensitive Web search challenging. First, scalability is of critical importance. With billions of publicly accessible documents, the Web is much larger than traditional datasets. Similarly, with millions of search queries issued each day, the query load is much higher than for traditional information retrieval systems. Second, there are no guarantees on the quality ofWeb pages, with Web-authors taking an adversarial, rather than cooperative, approach in attempts to inflate the rankings of their pages. Third, there is a significant amount of metadata embodied in the link structure corresponding to the hyperlinks between Web pages that can be exploitedduring the retrieval process. In this thesis, we design a search system, using the Stanford WebBase platform, that exploits the link structure of the Web to provide scalable, context-sensitive search.

3.0455156E-4 = product of:
  0.004568273 = sum of:
    0.004568273 = product of:
      0.009136546 = sum of:
        0.009136546 = weight(_text_:information in 4415) [ClassicSimilarity], result of:
          0.009136546 = score(doc=4415,freq=14.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.1796046 = fieldWeight in 4415, product of:
              3.7416575 = tf(freq=14.0), with freq of:
                14.0 = termFreq=14.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.02734375 = fieldNorm(doc=4415)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

This work proposes an approach which is intended to meet the particular challenges of Medical Language Processing, in particular medical information retrieval. At its core lies a new type of dictionary, in which the entries are equivalence classes of subwords, i.e., semantically minimal units. These equivalence classes capture intralingual as well as interlingual synonymy. As equivalence classes abstract away from subtle particularities within and between languages and reference to them is realized via a language-independent conceptual system, they form an interlingua. In this work, the theoretical foundations of this approach are elaborated on. Furthermore, design considerations of applications based on the subword methodology are drawn up and showcase implementations are evaluated in detail. Starting with the introduction of Medical Linguistics as a field of active research in Chapter two, its consideration as a domain separated form general linguistics is motivated. In particular, morphological phenomena inherent to medical language are figured in more detail, which leads to an alternative view on medical terms and the introduction of the notion of subwords. Chapter three describes the formal foundation of subwords and the underlying linguistic declarative as well as procedural knowledge. An implementation of the subword model for the medical domain, the MorphoSaurus system, is presented in Chapter four. Emphasis will be given on the multilingual aspect of the proposed approach, including English, German, and Portuguese. The automatic acquisition of (medical) subwords for other languages (Spanish, French, and Swedish), and their integration in already available resources is described in the fifth Chapter.
The proper handling of acronyms plays a crucial role in medical texts, e.g. in patient records, as well as in scientific literature. Chapter six presents an approach, in which acronyms are automatically acquired from (bio-) medical literature. Furthermore, acronyms and their definitions in different languages are linked to each other using the MorphoSaurus text processing system. Automatic word sense disambiguation is still one of the most challenging tasks in Natural Language Processing. In Chapter seven, cross-lingual considerations lead to a new methodology for automatic disambiguation applied to subwords. Beginning with Chapter eight, a series of applications based onMorphoSaurus are introduced. Firstly, the implementation of the subword approach within a crosslanguage information retrieval setting for the medical domain is described and evaluated on standard test document collections. In Chapter nine, this methodology is extended to multilingual information retrieval in the Web, for which user queries are translated into target languages based on the segmentation into subwords and their interlingual mappings. The cross-lingual, automatic assignment of document descriptors to documents is the topic of Chapter ten. A large-scale evaluation of a heuristic, as well as a statistical algorithm is carried out using a prominent medical thesaurus as a controlled vocabulary. In Chapter eleven, it will be shown how MorphoSaurus can be used to map monolingual, lexical resources across different languages. As a result, a large multilingual medical lexicon with high coverage and complete lexical information is built and evaluated against a comparable, already available and commonly used lexical repository for the medical domain. Chapter twelve sketches a few applications based on MorphoSaurus. The generality and applicability of the subword approach to other domains is outlined, and proof-of-concepts in real-world scenarios are presented. Finally, Chapter thirteen recapitulates the most important aspects of MorphoSaurus and the potential benefit of its employment in medical information systems is carefully assessed, both for medical experts in their everyday life, but also with regard to health care consumers and their existential information needs.

Source

Subword indexing, lexical learning and word sense disambiguation for medical crosslanguage information retrieval

2.941635E-4 = product of:
  0.004412452 = sum of:
    0.004412452 = product of:
      0.008824904 = sum of:
        0.008824904 = weight(_text_:information in 2191) [ClassicSimilarity], result of:
          0.008824904 = score(doc=2191,freq=10.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.1734784 = fieldWeight in 2191, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=2191)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

In the Internet information space, advanced information retrieval (IR) methods and automatic text processing are used in conjunction with traditional knowledge organization systems (KOS). New information technology provides a platform for better KOS publishing, exploitation and sharing both for human and machine use. Networked KOS services are now being planned and developed as powerful tools for resource discovery. They will enable automatic contextualisation, interpretation and query matching to different indexing languages. The Semantic Web promises to be an environment in which the quality of semantic relationships in bibliographic classification systems can be fully exploited. Their use in the networked environment is, however, limited by the fact that they are not prepared or made available for advanced machine processing. The UDC was chosen for this research because of its widespread use and its long-term presence in online information retrieval systems. It was also the first system to be used for the automatic classification of Internet resources, and the first to be made available as a classification tool on the Web. The objective of this research is to establish the advantages of using UDC for information retrieval in a networked environment, to highlight the problems of automation and classification exchange, and to offer possible solutions. The first research question was is there enough evidence of the use of classification on the Internet to justify further development with this particular environment in mind? The second question is what are the automation requirements for the full exploitation of UDC and its exchange? The third question is which areas are in need of improvement and what specific recommendations can be made for implementing the UDC in a networked environment? A summary of changes required in the management and development of the UDC to facilitate its full adaptation for future use is drawn from this analysis.

2.941635E-4 = product of:
  0.004412452 = sum of:
    0.004412452 = product of:
      0.008824904 = sum of:
        0.008824904 = weight(_text_:information in 535) [ClassicSimilarity], result of:
          0.008824904 = score(doc=535,freq=10.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.1734784 = fieldWeight in 535, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=535)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

Literature-based discovery (LBD) research aims at finding effective computational methods for predicting previously unknown connections between clusters of research papers from disparate research areas. Existing methods encompass two general approaches. The first approach searches for these unknown connections by examining the textual contents of research papers. In addition to the existing textual features, the second approach incorporates structural features of scientific literatures, such as citation structures. These approaches, however, have not considered research papers' latent bibliographic metadata structures as important features that can be used for predicting previously unknown relationships between them. This thesis investigates a new graph-based LBD method that exploits the latent bibliographic metadata connections between pairs of research papers. The heterogeneous bibliographic information network is proposed as an efficient graph-based data structure for modeling the complex relationships between these metadata. In contrast to previous approaches, this method seamlessly combines textual and citation information in the form of pathbased metadata features for predicting future co-citation links between research papers from disparate research fields. The results reported in this thesis provide evidence that the method is effective for reconstructing the historical literature-based discovery hypotheses. This thesis also investigates the effects of semantic modeling and topic modeling on the performance of the proposed method. For semantic modeling, a general-purpose word sense disambiguation technique is proposed to reduce the lexical ambiguity in the title and abstract of research papers. The experimental results suggest that the reduced lexical ambiguity did not necessarily lead to a better performance of the method. This thesis discusses some of the possible contributing factors to these results. Finally, topic modeling is used for learning the latent topical relations between research papers. The learned topic model is incorporated into the heterogeneous bibliographic information network graph and allows new predictive features to be learned. The results in this thesis suggest that topic modeling improves the performance of the proposed method by increasing the overall accuracy for predicting the future co-citation links between disparate research papers.

Footnote

A thesis submitted in ful llment of the requirements for the degree of Doctor of Philosophy Monash University, Faculty of Information Technology.

2.6310782E-4 = product of:
  0.0039466172 = sum of:
    0.0039466172 = product of:
      0.0078932345 = sum of:
        0.0078932345 = weight(_text_:information in 4569) [ClassicSimilarity], result of:
          0.0078932345 = score(doc=4569,freq=8.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.1551638 = fieldWeight in 4569, product of:
              2.828427 = tf(freq=8.0), with freq of:
                8.0 = termFreq=8.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=4569)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

The results of this thesis are intended to support the information architect in designing a solution for improved search in a corporate environment. Specifically we have examined the type of search problems that require a domain model to enhance the search process. There are several approaches to modeling a domain. We have considered 3 different types of domain modeling schemes; taxonomy, thesaurus and ontology. The intention is to support the information architect in making an informed choice between one or more of these schemes. In our opinion the main criteria for this choice are the modeling characteristics of a scheme and the suitability for application in the search process. The second chapter is a discussion of modeling characteristics of each scheme, followed by a comparison between them. This should give an information architect an idea of which aspects of a domain can be modeled with each scheme. What is missing here is an indication of the effort required to model a domain with each scheme. There are too many factors that influence the amount of required effort, ranging from measurable factors like domain size and resource characteristics to cultural matters such as the willingness to share knowledge and the existence of a project champion in the team to keep the project running. The third chapter shows what role domain models can play in each part of the search process. This gives an idea of the problems that domain models can solve. We have split the search process into individual parts to show that domain models can be applied very differently in the process. The fourth chapter makes recommendations about the suitability of each individualdomain modeling scheme for improving search. Each scheme has particular characteristics that make it especially suitable for a domain or a search problem. In the appendix each case study is described in detail. These descriptions are intended to serve as a benchmark. The current problem of the enterprise can be compared to those described to see which case study is most similar, which solution was chosen, which problems arose and how they were dealt with. An important issue that we have not touched upon in this thesis is that of maintenance. The real problems of a domain model are revealed when it is applied in a search system and its deficits and wrong assumptions become clear. Adaptation and maintenance are always required. Unfortunately we have not been able to glean sufficient information about maintenance issues from our case studies to draw any meaningful conclusions.

2.3255666E-4 = product of:
  0.0034883497 = sum of:
    0.0034883497 = product of:
      0.0069766995 = sum of:
        0.0069766995 = weight(_text_:information in 1210) [ClassicSimilarity], result of:
          0.0069766995 = score(doc=1210,freq=4.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.13714671 = fieldWeight in 1210, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.0390625 = fieldNorm(doc=1210)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

The current web consists of documents that are highly heterogeneous and hard for machines to understand. The Semantic Web is a progressive movement of the Word Wide Web, aiming at converting the current web of unstructured documents to the web of data. In the Semantic Web, web documents are annotated with metadata using standardized ontology language. These annotated documents are directly processable by machines and it highly improves their usability and usefulness. In Ericsson, similar problems occur. There are massive documents being created with well-defined structures. Though these documents are about domain specific knowledge and can have rich relations, they are currently managed by a traditional search engine, which ignores the rich domain specific information and presents few data to users. Motivated by the Semantic Web, we aim to find standard ways to process these documents, extract rich domain specific information and annotate these data to documents with formal markup languages. We propose this project to develop a domain specific search engine for processing different documents and building explicit relations for them. This research project consists of the three main focuses: examining different domain specific documents and finding ways to extract their metadata; integrating a text search engine with an ontology server; exploring novel ways to build relations for documents. We implement this system and demonstrate its functions. As a prototype, the system provides required features and will be extended in the future.

2.3255666E-4 = product of:
  0.0034883497 = sum of:
    0.0034883497 = product of:
      0.0069766995 = sum of:
        0.0069766995 = weight(_text_:information in 3465) [ClassicSimilarity], result of:
          0.0069766995 = score(doc=3465,freq=4.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.13714671 = fieldWeight in 3465, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.0390625 = fieldNorm(doc=3465)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

This dissertation discusses, explains, and demonstrates a new algorithm for translating chemica l nomenclature into molecular formulas. In order to place the study in its proper context and perspective the historical development of nomenclature is first discussed, aa well as other related aspects of the chemical information problem. The relationship of nomenclature to modern linguistic studies is then introduced. Tire relevance of structural linguistic procedures to the study of chemical nomenclature is shown. The methods of the linguist are illustrated by examples from chemical discourse. The algorithm is then explained, first for the human translator and then for use by a computer. Flow diagrams for the computer syntactic analysis, dictionary Iook-up routine, and furmula calculation routine are included. The sampling procedure for testing the algorithm is explained and finalIy, conclusions are drawn with respect to the general validity of the method and the dirsction that might be taken for future research. A summary of modern chemical nomenclature practice is appened primarily for use by the reader who is not familiar with chemical nomenclature.

Content

Doctoral dissertation, University of Pennsylvania, 1961. Vgl..: http://www.garfield.library.upenn.edu/essays/v7p441y1984.pdf. Auch in: Essays of an information scientist. Vol. 7. Philadelphia, PA: ISI Press, 1985. S.441-513.

2.3021935E-4 = product of:
  0.00345329 = sum of:
    0.00345329 = product of:
      0.00690658 = sum of:
        0.00690658 = weight(_text_:information in 1653) [ClassicSimilarity], result of:
          0.00690658 = score(doc=1653,freq=2.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.13576832 = fieldWeight in 1653, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.0546875 = fieldNorm(doc=1653)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

The thesis deals with the evaluation of mathematics information retrieval (IR). It gives an overview of the history of regular IR evaluation, initiatives that are engaged in this field of research as well as most common methods and measures used for evaluation. The findings are applied to the specifics of mathematics retrieval. This thesis also summarizes the state-of-the-art of MIaS math search system, which is already being used in an international web portal. Latest developments aiming towards the second version of the system are described. In addition to participating in the international evaluation conference and workshop, MIaS is tested for effectiveness and efficiency in this work. Measured performance indicators are evaluated and future work is suggested accordingly.

2.1815748E-4 = product of:
  0.003272362 = sum of:
    0.003272362 = product of:
      0.006544724 = sum of:
        0.006544724 = weight(_text_:information in 4472) [ClassicSimilarity], result of:
          0.006544724 = score(doc=4472,freq=22.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.12865502 = fieldWeight in 4472, product of:
              4.690416 = tf(freq=22.0), with freq of:
                22.0 = termFreq=22.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.015625 = fieldNorm(doc=4472)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

Dans d'autres cas, le profil de l'utilisateur peut être mal exploité pour extraire ou inférer ses nouveaux besoins en information. Ce problème est beaucoup plus accentué avec les requêtes ambigües. Lorsque plusieurs centres d'intérêt auxquels est liée une requête ambiguë sont identifiés dans le profil de l'utilisateur, le système se voit incapable de sélectionner les données pertinentes depuis ce profil pour répondre à la requête. Ceci a un impact direct sur la qualité des résultats fournis à cet utilisateur. Afin de remédier à quelques-unes de ces limitations, nous nous sommes intéressés dans ce cadre de cette thèse de recherche au développement de techniques destinées principalement à l'amélioration de la pertinence des résultats des SRIs actuels et à faciliter l'exploration de grandes collections de documents. Pour ce faire, nous proposons une solution basée sur un nouveau concept d'indexation et de recherche d'information appelé la projection multi-espaces. Cette proposition repose sur l'exploitation de différentes catégories d'information sémantiques et sociales qui permettent d'enrichir l'univers de représentation des documents et des requêtes de recherche en plusieurs dimensions d'interprétations. L'originalité de cette représentation est de pouvoir distinguer entre les différentes interprétations utilisées pour la description et la recherche des documents. Ceci donne une meilleure visibilité sur les résultats retournés et aide à apporter une meilleure flexibilité de recherche et d'exploration, en donnant à l'utilisateur la possibilité de naviguer une ou plusieurs vues de données qui l'intéressent le plus. En outre, les univers multidimensionnels de représentation proposés pour la description des documents et l'interprétation des requêtes de recherche aident à améliorer la pertinence des résultats de l'utilisateur en offrant une diversité de recherche/exploration qui aide à répondre à ses différents besoins et à ceux des autres différents utilisateurs. Cette étude exploite différents aspects liés à la recherche personnalisée et vise à résoudre les problèmes engendrés par l'évolution des besoins en information de l'utilisateur. Ainsi, lorsque le profil de cet utilisateur est utilisé par notre système, une technique est proposée et employée pour identifier les intérêts les plus représentatifs de ses besoins actuels dans son profil. Cette technique se base sur la combinaison de trois facteurs influents, notamment le facteur contextuel, fréquentiel et temporel des données. La capacité des utilisateurs à interagir, à échanger des idées et d'opinions, et à former des réseaux sociaux sur le Web, a amené les systèmes à s'intéresser aux types d'interactions de ces utilisateurs, au niveau d'interaction entre eux ainsi qu'à leurs rôles sociaux dans le système. Ces informations sociales sont abordées et intégrées dans ce travail de recherche. L'impact et la manière de leur intégration dans le processus de RI sont étudiés pour améliorer la pertinence des résultats.
Since its appearance in the early 90's, the World Wide Web (WWW or Web) has provided universal access to knowledge and the world of information has been primarily witness to a great revolution (the digital revolution). It quickly became very popular, making it the largest and most comprehensive database and knowledge base thanks to the amount and diversity of data it contains. However, the considerable increase and evolution of these data raises important problems for users, in particular for accessing the documents most relevant to their search queries. In order to cope with this exponential explosion of data volume and facilitate their access by users, various models are offered by information retrieval systems (IRS) for the representation and retrieval of web documents. Traditional SRIs use simple keywords that are not semantically linked to index and retrieve these documents. This creates limitations in terms of the relevance and ease of exploration of results. To overcome these limitations, existing techniques enrich documents by integrating external keywords from different sources. However, these systems still suffer from limitations that are related to the exploitation techniques of these sources of enrichment. When the different sources are used so that they cannot be distinguished by the system, this limits the flexibility of the exploration models that can be applied to the results returned by this system. Users then feel lost to these results, and find themselves forced to filter them manually to select the relevant information. If they want to go further, they must reformulate and target their search queries even more until they reach the documents that best meet their expectations. In this way, even if the systems manage to find more relevant results, their presentation remains problematic. In order to target research to more user-specific information needs and improve the relevance and exploration of its research findings, advanced SRIs adopt different data personalization techniques that assume that current research of user is directly related to his profile and / or previous browsing / search experiences.
However, this assumption does not hold in all cases, the needs of the user evolve over time and can move away from his previous interests stored in his profile. In other cases, the user's profile may be misused to extract or infer new information needs. This problem is much more accentuated with ambiguous queries. When multiple POIs linked to a search query are identified in the user's profile, the system is unable to select the relevant data from that profile to respond to that request. This has a direct impact on the quality of the results provided to this user. In order to overcome some of these limitations, in this research thesis, we have been interested in the development of techniques aimed mainly at improving the relevance of the results of current SRIs and facilitating the exploration of major collections of documents. To do this, we propose a solution based on a new concept and model of indexing and information retrieval called multi-spaces projection. This proposal is based on the exploitation of different categories of semantic and social information that enrich the universe of document representation and search queries in several dimensions of interpretations. The originality of this representation is to be able to distinguish between the different interpretations used for the description and the search for documents. This gives a better visibility on the results returned and helps to provide a greater flexibility of search and exploration, giving the user the ability to navigate one or more views of data that interest him the most. In addition, the proposed multidimensional representation universes for document description and search query interpretation help to improve the relevance of the user's results by providing a diversity of research / exploration that helps meet his diverse needs and those of other different users. This study exploits different aspects that are related to the personalized search and aims to solve the problems caused by the evolution of the information needs of the user. Thus, when the profile of this user is used by our system, a technique is proposed and used to identify the interests most representative of his current needs in his profile. This technique is based on the combination of three influential factors, including the contextual, frequency and temporal factor of the data. The ability of users to interact, exchange ideas and opinions, and form social networks on the Web, has led systems to focus on the types of interactions these users have at the level of interaction between them as well as their social roles in the system. This social information is discussed and integrated into this research work. The impact and how they are integrated into the IR process are studied to improve the relevance of the results.

2.1753683E-4 = product of:
  0.0032630523 = sum of:
    0.0032630523 = product of:
      0.0065261046 = sum of:
        0.0065261046 = weight(_text_:information in 4232) [ClassicSimilarity], result of:
          0.0065261046 = score(doc=4232,freq=14.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.128289 = fieldWeight in 4232, product of:
              3.7416575 = tf(freq=14.0), with freq of:
                14.0 = termFreq=14.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.01953125 = fieldNorm(doc=4232)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

After the launch of the World Wide Web, it became clear that searching documentson the Web would not be trivial. Well-known engines to search the web, like Google, focus on search in web documents using keywords. The documents are structured and indexed to ensure keywords match documents as accurately as possible. However, searching by keywords does not always suice. It is oen the case that users do not know exactly how to formulate the search query or which keywords guarantee retrieving the most relevant documents. Besides that, it occurs that users rather want to browse information than looking up something specific. It turned out that there is need for systems that enable more interactivity and facilitate the gradual refinement of search queries to explore the Web. Users expect more from the Web because the short keyword-based queries they pose during search, do not suffice for all cases. On top of that, the Web is changing structurally. The Web comprises, apart from a collection of documents, more and more linked data, pieces of information structured so they can be processed by machines. The consequently applied semantics allow users to exactly indicate machines their search intentions. This is made possible by describing data following controlled vocabularies, concept lists composed by experts, published uniquely identifiable on the Web. Even so, it is still not trivial to explore data on the Web. There is a large variety of vocabularies and various data sources use different terms to identify the same concepts.
This PhD-thesis describes how to effectively explore linked data on the Web. The main focus is on scenarios where users want to discover relationships between resources rather than finding out more about something specific. Searching for a specific document or piece of information fits in the theoretical framework of information retrieval and is associated with exploratory search. Exploratory search goes beyond 'looking up something' when users are seeking more detailed understanding, further investigation or navigation of the initial search results. The ideas behind exploratory search and querying linked data merge when it comes to the way knowledge is represented and indexed by machines - how data is structured and stored for optimal searchability. Queries and information should be aligned to facilitate that searches also reveal connections between results. This implies that they take into account the same semantic entities, relevant at that moment. To realize this, we research three techniques that are evaluated one by one in an experimental set-up to assess how well they succeed in their goals. In the end, the techniques are applied to a practical use case that focuses on forming a bridge between the Web and the use of digital libraries in scientific research. Our first technique focuses on the interactive visualization of search results. Linked data resources can be brought in relation with each other at will. This leads to complex and diverse graphs structures. Our technique facilitates navigation and supports a workflow starting from a broad overview on the data and allows narrowing down until the desired level of detail to then broaden again. To validate the flow, two visualizations where implemented and presented to test-users. The users judged the usability of the visualizations, how the visualizations fit in the workflow and to which degree their features seemed useful for the exploration of linked data.
The ideas behind exploratory search and querying linked data merge when it comes to the way knowledge is represented and indexed by machines - how data is structured and stored for optimal searchability. eries and information should be aligned to facilitate that searches also reveal connections between results. This implies that they take into account the same semantic entities, relevant at that moment. To realize this, we research three techniques that are evaluated one by one in an experimental set-up to assess how well they succeed in their goals. In the end, the techniques are applied to a practical use case that focuses on forming a bridge between the Web and the use of digital libraries in scientific research.
When we speak about finding relationships between resources, it is necessary to dive deeper in the structure. The graph structure of linked data where the semantics give meaning to the relationships between resources enable the execution of pathfinding algorithms. The assigned weights and heuristics are base components of such algorithms and ultimately define (the order) which resources are included in a path. These paths explain indirect connections between resources. Our third technique proposes an algorithm that optimizes the choice of resources in terms of serendipity. Some optimizations guard the consistence of candidate-paths where the coherence of consecutive connections is maximized to avoid trivial and too arbitrary paths. The implementation uses the A* algorithm, the de-facto reference when it comes to heuristically optimized minimal cost paths. The effectiveness of paths was measured based on common automatic metrics and surveys where the users could indicate their preference for paths, generated each time in a different way. Finally, all our techniques are applied to a use case about publications in digital libraries where they are aligned with information about scientific conferences and researchers. The application to this use case is a practical example because the different aspects of exploratory search come together. In fact, the techniques also evolved from the experiences when implementing the use case. Practical details about the semantic model are explained and the implementation of the search system is clarified module by module. The evaluation positions the result, a prototype of a tool to explore scientific publications, researchers and conferences next to some important alternatives.

1.8604532E-4 = product of:
  0.0027906797 = sum of:
    0.0027906797 = product of:
      0.0055813594 = sum of:
        0.0055813594 = weight(_text_:information in 4746) [ClassicSimilarity], result of:
          0.0055813594 = score(doc=4746,freq=4.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.10971737 = fieldWeight in 4746, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=4746)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

Many real-world domains can be represented as large node-link graphs: backbone Internet routers connect with 70,000 other hosts, mid-sized Web servers handle between 20,000 and 200,000 hyperlinked documents, and dictionaries contain millions of words defined in terms of each other. Computational manipulation of such large graphs is common, but previous tools for graph visualization have been limited to datasets of a few thousand nodes. Visual depictions of graphs and networks are external representations that exploit human visual processing to reduce the cognitive load of many tasks that require understanding of global or local structure. We assert that the two key advantages of computer-based systems for information visualization over traditional paper-based visual exposition are interactivity and scalability. We also argue that designing visualization software by taking the characteristics of a target user's task domain into account leads to systems that are more effective and scale to larger datasets than previous work. This thesis contains a detailed analysis of three specialized systems for the interactive exploration of large graphs, relating the intended tasks to the spatial layout and visual encoding choices. We present two novel algorithms for specialized layout and drawing that use quite different visual metaphors. The H3 system for visualizing the hyperlink structures of web sites scales to datasets of over 100,000 nodes by using a carefully chosen spanning tree as the layout backbone, 3D hyperbolic geometry for a Focus+Context view, and provides a fluid interactive experience through guaranteed frame rate drawing. The Constellation system features a highly specialized 2D layout intended to spatially encode domain-specific information for computational linguists checking the plausibility of a large semantic network created from dictionaries. The Planet Multicast system for displaying the tunnel topology of the Internet's multicast backbone provides a literal 3D geographic layout of arcs on a globe to help MBone maintainers find misconfigured long-distance tunnels. Each of these three systems provides a very different view of the graph structure, and we evaluate their efficacy for the intended task. We generalize these findings in our analysis of the importance of interactivity and specialization for graph visualization systems that are effective and scalable.

1.8604532E-4 = product of:
  0.0027906797 = sum of:
    0.0027906797 = product of:
      0.0055813594 = sum of:
        0.0055813594 = weight(_text_:information in 5819) [ClassicSimilarity], result of:
          0.0055813594 = score(doc=5819,freq=4.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.10971737 = fieldWeight in 5819, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.03125 = fieldNorm(doc=5819)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

Industries have adopted a standardized set of practices for developing their products. In the automotive domain, the provision of safety-compliant systems is guided by ISO 26262, a standard that specifies a set of requirements and recommendations for developing automotive safety-critical systems. For being in compliance with ISO 26262, the safety lifecycle proposed by the standard must be included in the development process of a vehicle. Besides, a safety case that shows that the system is acceptably safe has to be provided. The provision of a safety case implies the execution of a precise documentation process. This process makes sure that the work products are available and traceable. Further, the documentation management is defined in the standard as a mandatory activity and guidelines are proposed/imposed for its elaboration. It would be appropriate to point out that a well-documented safety lifecycle will provide the necessary inputs for the generation of an ISO 26262-compliant safety case. The OSLC (Open Services for Lifecycle Collaboration) standard and the maturing stack of semantic web technologies represent a promising integration platform for enabling semantic interoperability between the tools involved in the safety lifecycle. Tools for requirements, architecture, development management, among others, are expected to interact and shared data with the help of domains specifications created in OSLC. This thesis proposes the creation of an OSLC tool-chain infrastructure for sharing safety-related information, where fragments of safety information can be generated. The steps carried out during the elaboration of this master thesis consist in the identification, representation, and shaping of the RDF resources needed for the creation of a safety case. The focus of the thesis is limited to a tiny portion of the ISO 26262 left-hand side of the V-model, more exactly part 6 clause 8 of the standard: Software unit design and implementation. Regardless of the use of a restricted portion of the standard during the execution of this thesis, the findings can be extended to other parts, and the conclusions can be generalize. This master thesis is considered one of the first steps towards the provision of an OSLC-based and ISO 26262-compliant methodological approach for representing and shaping the work products resulting from the execution of the safety lifecycle, documentation required in the conformation of an ISO-compliant safety case.

1.1510967E-4 = product of:
  0.001726645 = sum of:
    0.001726645 = product of:
      0.00345329 = sum of:
        0.00345329 = weight(_text_:information in 1536) [ClassicSimilarity], result of:
          0.00345329 = score(doc=1536,freq=2.0), product of:
            0.050870337 = queryWeight, product of:
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.028978055 = queryNorm
            0.06788416 = fieldWeight in 1536, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              1.7554779 = idf(docFreq=20772, maxDocs=44218)
              0.02734375 = fieldNorm(doc=1536)
      0.5 = coord(1/2)
  0.06666667 = coord(1/15)

Abstract

Multiword expressions (MWEs) are lexical items that can be decomposed into single words and display lexical, syntactic, semantic, pragmatic and/or statistical idiosyncrasy (Sag et al., 2002; Kim, 2008; Calzolari et al., 2002). The proper treatment of multiword expressions such as rock 'n' roll and make a decision is essential for many natural language processing (NLP) applications like information extraction and retrieval, terminology extraction and machine translation, and it is important to identify multiword expressions in context. For example, in machine translation we must know that MWEs form one semantic unit, hence their parts should not be translated separately. For this, multiword expressions should be identified first in the text to be translated. The chief aim of this thesis is to develop machine learning-based approaches for the automatic detection of different types of multiword expressions in English and Hungarian natural language texts. In our investigations, we pay attention to the characteristics of different types of multiword expressions such as nominal compounds, multiword named entities and light verb constructions, and we apply novel methods to identify MWEs in raw texts. In the thesis it will be demonstrated that nominal compounds and multiword amed entities may require a similar approach for their automatic detection as they behave in the same way from a linguistic point of view. Furthermore, it will be shown that the automatic detection of light verb constructions can be carried out using two effective machine learning-based approaches.