Search (16 results, page 1 of 1)

0.0503529 = product of:
  0.1510587 = sum of:
    0.1510587 = sum of:
      0.06755122 = weight(_text_:system in 100) [ClassicSimilarity], result of:
        0.06755122 = score(doc=100,freq=2.0), product of:
          0.16177002 = queryWeight, product of:
            3.1495528 = idf(docFreq=5152, maxDocs=44218)
            0.051362853 = queryNorm
          0.41757566 = fieldWeight in 100, product of:
            1.4142135 = tf(freq=2.0), with freq of:
              2.0 = termFreq=2.0
            3.1495528 = idf(docFreq=5152, maxDocs=44218)
            0.09375 = fieldNorm(doc=100)
      0.083507486 = weight(_text_:22 in 100) [ClassicSimilarity], result of:
        0.083507486 = score(doc=100,freq=2.0), product of:
          0.17986396 = queryWeight, product of:
            3.5018296 = idf(docFreq=3622, maxDocs=44218)
            0.051362853 = queryNorm
          0.46428138 = fieldWeight in 100, product of:
            1.4142135 = tf(freq=2.0), with freq of:
              2.0 = termFreq=2.0
            3.5018296 = idf(docFreq=3622, maxDocs=44218)
            0.09375 = fieldNorm(doc=100)
  0.33333334 = coord(1/3)

Date

22. 9.2007 15:41:14

0.013788836 = product of:
  0.041366506 = sum of:
    0.041366506 = product of:
      0.08273301 = sum of:
        0.08273301 = weight(_text_:system in 202) [ClassicSimilarity], result of:
          0.08273301 = score(doc=202,freq=12.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.51142365 = fieldWeight in 202, product of:
              3.4641016 = tf(freq=12.0), with freq of:
                12.0 = termFreq=12.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.046875 = fieldNorm(doc=202)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

The authors present a model of information searching in thesaurus-enhanced search systems, intended as a reference model for system developers. The model focuses on user-system interaction and charts the specific stages of searching an indexed collection with a thesaurus. It was developed based on literature, findings from empirical studies, and analysis of existing systems. The model describes in detail the entities, processes, and decisions when interacting with a search system augmented with a thesaurus. A basic search scenario illustrates this process through the model. Graphical and textual depictions of the model are complemented by a concise matrix representation for evaluation purposes. Potential problems at different stages of the search process are discussed, together with possibilities for system developers. The aim is to set out a framework of processes, decisions, and risks involved in thesaurus-based search, within which system developers can consider potential avenues for support.

0.01289442 = product of:
  0.038683258 = sum of:
    0.038683258 = weight(_text_:management in 155) [ClassicSimilarity], result of:
      0.038683258 = score(doc=155,freq=2.0), product of:
        0.17312427 = queryWeight, product of:
          3.3706124 = idf(docFreq=4130, maxDocs=44218)
          0.051362853 = queryNorm
        0.22344214 = fieldWeight in 155, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          3.3706124 = idf(docFreq=4130, maxDocs=44218)
          0.046875 = fieldNorm(doc=155)
  0.33333334 = coord(1/3)

Source

Information processing and management. 33(1997) no.2, S.233-242

0.011598263 = product of:
  0.03479479 = sum of:
    0.03479479 = product of:
      0.06958958 = sum of:
        0.06958958 = weight(_text_:22 in 539) [ClassicSimilarity], result of:
          0.06958958 = score(doc=539,freq=2.0), product of:
            0.17986396 = queryWeight, product of:
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.051362853 = queryNorm
            0.38690117 = fieldWeight in 539, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.078125 = fieldNorm(doc=539)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Date

26.12.2011 13:22:07

0.0069589573 = product of:
  0.020876871 = sum of:
    0.020876871 = product of:
      0.041753743 = sum of:
        0.041753743 = weight(_text_:22 in 1347) [ClassicSimilarity], result of:
          0.041753743 = score(doc=1347,freq=2.0), product of:
            0.17986396 = queryWeight, product of:
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.051362853 = queryNorm
            0.23214069 = fieldWeight in 1347, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.046875 = fieldNorm(doc=1347)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Date

22. 8.2014 17:12:54

0.0066341567 = product of:
  0.01990247 = sum of:
    0.01990247 = product of:
      0.03980494 = sum of:
        0.03980494 = weight(_text_:system in 2703) [ClassicSimilarity], result of:
          0.03980494 = score(doc=2703,freq=4.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.24605882 = fieldWeight in 2703, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=2703)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

This paper discusses two inter-related themes: the retrieval potential of faceted thesauri and XML representations of fundamental facets. Initial findings are discussed from the ongoing 'FACET' project, in collaboration with the National Museum of Science and Industry. The work discussed seeks to take advantage of the structure afforded by faceted systems for multi-term queries and flexible matching, focusing in this paper an the Art and Architecture Thesaurus. A multi-term matching function yields ranked results with partial matches via semantic term expansion, based an a measure of distance over the semantic index space formed by thesaurus relationships. Our intention is to drive the system from general representations and a common query structure and interface. To this end, we are developing an XML representation based an work by the Classification Research Group an fundamental facets or categories. The XML representation maps categories to particular thesauri and hierarchies. The system interface, which is configured by the mapping, incorporates a thesaurus browser with navigation history together with a term search facility and drag and drop query builder.

0.0066341567 = product of:
  0.01990247 = sum of:
    0.01990247 = product of:
      0.03980494 = sum of:
        0.03980494 = weight(_text_:system in 4171) [ClassicSimilarity], result of:
          0.03980494 = score(doc=4171,freq=4.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.24605882 = fieldWeight in 4171, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=4171)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

Purpose - Traditional subject indexing and classification are considered infeasible in many digital collections. This paper seeks to investigate ways of enhancing social tagging via knowledge organization systems, with a view to improving the quality of tags for increased information discovery and retrieval performance. Design/methodology/approach - Enhanced tagging interfaces were developed for exemplar online repositories, and trials were undertaken with author and reader groups to evaluate the effectiveness of tagging augmented with control vocabulary for subject indexing of papers in online repositories. Findings - The results showed that using a knowledge organisation system to augment tagging does appear to increase the effectiveness of non-specialist users (that is, without information science training) in subject indexing. Research limitations/implications - While limited by the size and scope of the trials undertaken, these results do point to the usefulness of a mixed approach in supporting the subject indexing of online resources. Originality/value - The value of this work is as a guide to future developments in the practical support for resource indexing in online repositories.

0.0066341567 = product of:
  0.01990247 = sum of:
    0.01990247 = product of:
      0.03980494 = sum of:
        0.03980494 = weight(_text_:system in 600) [ClassicSimilarity], result of:
          0.03980494 = score(doc=600,freq=4.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.24605882 = fieldWeight in 600, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=600)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

Purpose The Integrative Levels Classification (ILC) is a comprehensive "freely faceted" knowledge organization system not previously expressed as SKOS (Simple Knowledge Organization System). This paper reports and reflects on work converting the ILC to SKOS representation. Design/methodology/approach The design of the ILC representation and the various steps in the conversion to SKOS are described and located within the context of previous work considering the representation of complex classification schemes in SKOS. Various issues and trade-offs emerging from the conversion are discussed. The conversion implementation employed the STELETO transformation tool. Findings The ILC conversion captures some of the ILC facet structure by a limited extension beyond the SKOS standard. SPARQL examples illustrate how this extension could be used to create faceted, compound descriptors when indexing or cataloguing. Basic query patterns are provided that might underpin search systems. Possible routes for reducing complexity are discussed. Originality/value Complex classification schemes, such as the ILC, have features which are not straight forward to represent in SKOS and which extend beyond the functionality of the SKOS standard. The ILC's facet indicators are modelled as rdf:Property sub-hierarchies that accompany the SKOS RDF statements. The ILC's top-level fundamental facet relationships are modelled by extensions of the associative relationship - specialised sub-properties of skos:related. An approach for representing faceted compound descriptions in ILC and other faceted classification schemes is proposed.

0.0065674796 = product of:
  0.019702438 = sum of:
    0.019702438 = product of:
      0.039404877 = sum of:
        0.039404877 = weight(_text_:system in 2801) [ClassicSimilarity], result of:
          0.039404877 = score(doc=2801,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.2435858 = fieldWeight in 2801, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0546875 = fieldNorm(doc=2801)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

Reviews claims made for virtual architectures and proposes a semantic data model for hypermedia architecture. Semantic modelling and an extended binary relational model in particular, are outlined in the context of hypermedia. The binary relational store is a simple, uniform data structure, capable of representing abstraction in the application model. Pilot implementations of museum hypermedia systems demonstrate that the architecture is capable of supporting a variety of navigation techniques and authoring tools. Outlines the SHIC (Social History and Industrial Classification) museum classification schema, and discusses its implementation in a hypermedia system based on a binary relational store. Considers experiences with the prototypes and discusses feedback from the museum profession and general public. An extended binary relational model is particularly suited to certain forms of reasoning based on generalization

0.0056292685 = product of:
  0.016887804 = sum of:
    0.016887804 = product of:
      0.03377561 = sum of:
        0.03377561 = weight(_text_:system in 7460) [ClassicSimilarity], result of:
          0.03377561 = score(doc=7460,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.20878783 = fieldWeight in 7460, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.046875 = fieldNorm(doc=7460)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

Describes an integrated approach to similarity coefficients for information spaces with multiple dimensions of different types of index term. Categorises applications of similarity coefficients underlying different navigation tools in hypermedia by type of term. Describes an implementation of a unified similarity coefficient based on work in numerical taxonomy, with illustrative scenarios from an experimental navigation via similarity tool for a prototype social history museum hypermedia system. The underlying architecture is based on a semantic approach, where semantic relationships can exist between index terms. This allows imprecise matching when comparing for similarity, with distance measures yielding a degree of match. A ranked list of matching items over several weighted dimensions is returned by the similarity navigation tool. The approach has the potential of allowing different access methods to multimedia data to be combined

0.004691057 = product of:
  0.01407317 = sum of:
    0.01407317 = product of:
      0.02814634 = sum of:
        0.02814634 = weight(_text_:system in 2215) [ClassicSimilarity], result of:
          0.02814634 = score(doc=2215,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.17398985 = fieldWeight in 2215, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=2215)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

Purpose - The purpose of this paper is to explore query expansion via conceptual distance in thesaurus indexed collections Design/methodology/approach - An extract of the National Museum of Science and Industry's collections database, indexed with the Getty Art and Architecture Thesaurus (AAT), was the dataset for the research. The system architecture and algorithms for semantic closeness and the matching function are outlined. Standalone and web interfaces are described and formative qualitative user studies are discussed. One user session is discussed in detail, together with a scenario based on a related public inquiry. Findings are set in context of the literature on thesaurus-based query expansion. This paper discusses the potential of query expansion techniques using the semantic relationships in a faceted thesaurus. Findings - Thesaurus-assisted retrieval systems have potential for multi-concept descriptors, permitting very precise queries and indexing. However, indexer and searcher may differ in terminology judgments and there may not be any exactly matching results. The integration of semantic closeness in the matching function permits ranked results for multi-concept queries in thesaurus-indexed applications. An in-memory representation of the thesaurus semantic network allows a combination of automatic and interactive control of expansion and control of expansion on individual query terms. Originality/value - The application of semantic expansion to browsing may be useful in interface options where thesaurus structure is hidden.

0.004691057 = product of:
  0.01407317 = sum of:
    0.01407317 = product of:
      0.02814634 = sum of:
        0.02814634 = weight(_text_:system in 2918) [ClassicSimilarity], result of:
          0.02814634 = score(doc=2918,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.17398985 = fieldWeight in 2918, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=2918)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

Purpose - The purpose of this paper is to explore the potential of applying the Dewey Decimal Classification (DDC) as an established knowledge organization system (KOS) for enhancing social tagging, with the ultimate purpose of improving subject indexing and information retrieval. Design/methodology/approach - Over 11.000 Intute metadata records in politics were used. Totally, 28 politics students were each given four tasks, in which a total of 60 resources were tagged in two different configurations, one with uncontrolled social tags only and another with uncontrolled social tags as well as suggestions from a controlled vocabulary. The controlled vocabulary was DDC comprising also mappings from the Library of Congress Subject Headings. Findings - The results demonstrate the importance of controlled vocabulary suggestions for indexing and retrieval: to help produce ideas of which tags to use, to make it easier to find focus for the tagging, to ensure consistency and to increase the number of access points in retrieval. The value and usefulness of the suggestions proved to be dependent on the quality of the suggestions, both as to conceptual relevance to the user and as to appropriateness of the terminology. Originality/value - No research has investigated the enhancement of social tagging with suggestions from the DDC, an established KOS, in a user trial, comparing social tagging only and social tagging enhanced with the suggestions. This paper is a final reflection on all aspects of the study.

0.004691057 = product of:
  0.01407317 = sum of:
    0.01407317 = product of:
      0.02814634 = sum of:
        0.02814634 = weight(_text_:system in 3179) [ClassicSimilarity], result of:
          0.02814634 = score(doc=3179,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.17398985 = fieldWeight in 3179, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=3179)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

There are many advantages for Digital Libraries in indexing with classifications or thesauri, but some current disincentive in the lack of flexible retrieval tools that deal with compound descriptors. This paper discusses a matching function for compound descriptors, or multi-concept subject headings, that does not rely an exact matching but incorporates term expansion via thesaurus semantic relationships to produce ranked results that take account of missing and partially matching terms. The matching function is based an a measure of semantic closeness between terms, which has the potential to help with recall problems. The work reported is part of the ongoing FACET project in collaboration with the National Museum of Science and Industry and its collections database. The architecture of the prototype system and its Interface are outlined. The matching problem for compound descriptors is reviewed and the FACET implementation described. Results are discussed from scenarios using the faceted Getty Art and Architecture Thesaurus. We argue that automatic traversal of thesaurus relationships can augment the user's browsing possibilities. The techniques can be applied both to unstructured multi-concept subject headings and potentially to more syntactically structured strings. The notion of a focus term is used by the matching function to model AAT modified descriptors (noun phrases). The relevance of the approach to precoordinated indexing and matching faceted strings is discussed.

0.004691057 = product of:
  0.01407317 = sum of:
    0.01407317 = product of:
      0.02814634 = sum of:
        0.02814634 = weight(_text_:system in 2895) [ClassicSimilarity], result of:
          0.02814634 = score(doc=2895,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.17398985 = fieldWeight in 2895, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.0390625 = fieldNorm(doc=2895)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

The article presents a method for automatic semantic indexing of archaeological grey-literature reports using empirical (rule-based) Information Extraction techniques in combination with domain-specific knowledge organization systems. The semantic annotation system (OPTIMA) performs the tasks of Named Entity Recognition, Relation Extraction, Negation Detection, and Word-Sense Disambiguation using hand-crafted rules and terminological resources for associating contextual abstractions with classes of the standard ontology CIDOC Conceptual Reference Model (CRM) for cultural heritage and its archaeological extension, CRM-EH. Relation Extraction (RE) performance benefits from a syntactic-based definition of RE patterns derived from domain oriented corpus analysis. The evaluation also shows clear benefit in the use of assistive natural language processing (NLP) modules relating to Word-Sense Disambiguation, Negation Detection, and Noun Phrase Validation, together with controlled thesaurus expansion. The semantic indexing results demonstrate the capacity of rule-based Information Extraction techniques to deliver interoperable semantic abstractions (semantic annotations) with respect to the CIDOC CRM and archaeological thesauri. Major contributions include recognition of relevant entities using shallow parsing NLP techniques driven by a complimentary use of ontological and terminological domain resources and empirical derivation of context-driven RE rules for the recognition of semantic relationships from phrases of unstructured text.

0.0046439096 = product of:
  0.013931728 = sum of:
    0.013931728 = product of:
      0.027863456 = sum of:
        0.027863456 = weight(_text_:system in 2677) [ClassicSimilarity], result of:
          0.027863456 = score(doc=2677,freq=4.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.17224117 = fieldWeight in 2677, product of:
              2.0 = tf(freq=4.0), with freq of:
                4.0 = termFreq=4.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.02734375 = fieldNorm(doc=2677)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

This paper discusses research that employs methods from Natural Language Processing (NLP) in exploiting the intellectual resources of Knowledge Organization Systems (KOS), particularly in the retrieval of information. A technique for the disambiguation of homographs and nominal compounds in free text, where these are known ambiguous terms in the KOS itself, is described. The use of Roget's Thesaurus as an intermediary in the process is also reported. A short review of the relevant literature in the field is given. Design considerations, results and conclusions are presented from the implementation of a prototype system. The linguistic techniques are applied at two complementary levels, namely an a free text string used as an entry point to the KOS, and an the underlying controlled vocabulary itself.

Content

1. Introduction The need for research into the application of linguistic techniques in Information Retrieval (IR) in general, and a similar need in faceted Knowledge Organization Systems (KOS) has been indicated by various authors. Smeaton (1997) points out the inherent limitations of conventional approaches to IR based an "bags of words", mainly difficulties caused by lexical ambiguity in the words concerned, and goes an to suggest the possibility of using Natural Language Processing (NLP) in query formulation. Past experience with a faceted retrieval system highlighted the need for integrating the linguistic perspective in order to fully utilise the potential of a KOS (Tudhope et al." 2002). The present research seeks to address some of these needs in using NLP to improve the efficacy of KOS tools in query and retrieval systems. Syntactic parsing and part-of-speech tagging can substantially reduce lexical ambiguity through homograph disambiguation. Given the two strings "1 fable the motion" and "I put the motion an the fable", for instance, the parser used in this research clearly indicates that 'fable' in the first string is a verb, while 'table' in the second string is a noun, a distinction that would be missed in the "bag of words" approach. This syntactic disambiguation enables a more precise matching from free text to the controlled vocabulary of a KOS and vice versa. The use of a general linguistic resource, namely Roget's Thesaurus of English Words and Phrases (RTEWP), as an intermediary in this process, is investigated. The adaptation of the Link parser (Sleator & Temperley, 1993) to the purposes of the research is reported. The design and implementation of the early practical stages of the project are described, and the results of the initial experiments are presented and evaluated. Applications of the techniques developed are foreseen in the areas of query disambiguation, information retrieval and automatic indexing. In the first section of the paper a brief review of the literature and relevant current work in the field is presented. The second section includes reports an the development of algorithms, the construction of data sets and theoretical and experimental work undertaken to date. The third section evaluates the results obtained, and outlines directions for future research.

0.0037528453 = product of:
  0.011258536 = sum of:
    0.011258536 = product of:
      0.022517072 = sum of:
        0.022517072 = weight(_text_:system in 175) [ClassicSimilarity], result of:
          0.022517072 = score(doc=175,freq=2.0), product of:
            0.16177002 = queryWeight, product of:
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.051362853 = queryNorm
            0.13919188 = fieldWeight in 175, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.1495528 = idf(docFreq=5152, maxDocs=44218)
              0.03125 = fieldNorm(doc=175)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Abstract

There are many advantages for Digital Libraries in indexing with classifications or thesauri, but some current disincentive in the lack of flexible retrieval tools that deal with compound descriptors. This demonstration of a research prototype illustrates a matching function for compound descriptors, or multi-concept subject headings, that does not rely on exact matching but incorporates term expansion via thesaurus semantic relationships to produce ranked results that take account of missing and partially matching terms. The matching function is based on a measure of semantic closeness between terms.The work is part of the EPSRC funded FACET project in collaboration with the UK National Museum of Science and Industry (NMSI) which includes the National Railway Museum. An export of NMSI's Collections Database is used as the dataset for the research. The J. Paul Getty Trust's Art and Architecture Thesaurus (AAT) is the main thesaurus in the project. The AAT is a widely used thesaurus (over 120,000 terms). Descriptors are organised in 7 facets representing separate conceptual classes of terms.The FACET application is a multi tiered architecture accessing a SQL Server database, with an OLE DB connection. The thesauri are stored as relational tables in the Server's database. However, a key component of the system is a parallel representation of the underlying semantic network as an in-memory structure of thesaurus concepts (corresponding to preferred terms). The structure models the hierarchical and associative interrelationships of thesaurus concepts via weighted poly-hierarchical links. Its primary purpose is real-time semantic expansion of query terms, achieved by a spreading activation semantic closeness algorithm. Queries with associated results are stored persistently using XML format data. A Visual Basic interface combines a thesaurus browser and an initial term search facility that takes into account equivalence relationships. Terms are dragged to a direct manipulation Query Builder which maintains the facet structure.