Search (25 results, page 1 of 2)

0.0139165055 = product of:
  0.041749515 = sum of:
    0.041749515 = weight(_text_:on in 3593) [ClassicSimilarity], result of:
      0.041749515 = score(doc=3593,freq=10.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.38036036 = fieldWeight in 3593, product of:
          3.1622777 = tf(freq=10.0), with freq of:
            10.0 = termFreq=10.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=3593)
  0.33333334 = coord(1/3)

Abstract

Considers the interindexer consistency between indexers working in various organisations and reports on the result of an empirical study. The interindexer consistency was low, but there were clear differences depending on whether the consistency was calculated on the basis to terms or concepts or aspects. The fact that the consistency figures remained low can be explained. The low indexing consistency caused by indexing errors also seems to be difficult to control. Indexing consistency and its control have a clear impact on how feasible and useful centralised services and union catalogues are and can be from the point of view of subject description.

Source

International forum on information and documentation. 15(1990) no.2, S.8-15

0.0124473 = product of:
  0.0373419 = sum of:
    0.0373419 = weight(_text_:on in 4028) [ClassicSimilarity], result of:
      0.0373419 = score(doc=4028,freq=8.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.34020463 = fieldWeight in 4028, product of:
          2.828427 = tf(freq=8.0), with freq of:
            8.0 = termFreq=8.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=4028)
  0.33333334 = coord(1/3)

Abstract

Reports on an analysis of lexical ambiguity in information retrieval text collections and on experiments to determine the utility of word meanings for separating relevant from nonrelevant documents. Results show that there is considerable ambiguity even in a specialised database. Word senses provide a significant separation between relevant and nonrelevant documents, but several factors contribute to determining whether disambiguation will make an improvement in performance such as: resolving lexical ambiguity was found to have little impact on retrieval effectiveness for documents that have many words in common with the query. Discusses other uses of word sense disambiguation in an information retrieval context

Source

ACM transactions on information systems. 10(1992) no.2, S.115-141

0.0124473 = product of:
  0.0373419 = sum of:
    0.0373419 = weight(_text_:on in 6941) [ClassicSimilarity], result of:
      0.0373419 = score(doc=6941,freq=8.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.34020463 = fieldWeight in 6941, product of:
          2.828427 = tf(freq=8.0), with freq of:
            8.0 = termFreq=8.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=6941)
  0.33333334 = coord(1/3)

Abstract

Explores the question of whether academic libraries keep up with the changes in the LCSH system. Analysis of the handling of 15 subject headings in 50 academic library catalogues available via the Internet found that libraries are not consistently maintaining subject authority control, or making syndetic references and scope notes in their catalogues. Discusses the results from the perspective of the libraries' performance, performance on the headings overall, performance on references, performance on the type of change made to the headings,a nd performance within 3 widely used onlien catalogue systems (DRA, INNOPAC and NOTIS). Discusses the implications of the findings in relationship to expressions of dissatisfaction with the effectiveness of subject cataloguing expressed by discussion groups on the Internet

0.0124473 = product of:
  0.0373419 = sum of:
    0.0373419 = weight(_text_:on in 579) [ClassicSimilarity], result of:
      0.0373419 = score(doc=579,freq=8.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.34020463 = fieldWeight in 579, product of:
          2.828427 = tf(freq=8.0), with freq of:
            8.0 = termFreq=8.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=579)
  0.33333334 = coord(1/3)

Abstract

This article presents a logical analysis of the characteristics of indexing and their effects on retrieval performance.It establishes the ability to ask the questions one needs to ask as the foundation of performance evaluation, and recall and discrimination as the basic quantitative performance measures for binary noninteractive retrieval systems. It then defines the characteristics of indexing that affect retrieval - namely, indexing devices, viewpoint-based and importance-based indexing exhaustivity, indexing specifity, indexing correctness, and indexing consistency - and examines in detail their effects on retrieval. It concludes that retrieval performance depends chiefly on the match between indexing and the requirements of the individual query and on the adaption of the query formulation to the characteristics of the retrieval system, and that the ensuing complexity must be considered in the design and testing of retrieval systems

0.012319634 = product of:
  0.0369589 = sum of:
    0.0369589 = weight(_text_:on in 4779) [ClassicSimilarity], result of:
      0.0369589 = score(doc=4779,freq=6.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.33671528 = fieldWeight in 4779, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0625 = fieldNorm(doc=4779)
  0.33333334 = coord(1/3)

Abstract

The interindexer consistency between indexers working in 10 libraries was considered. The indexing environment is described with the help of organisational theory. Interindexer consistency was low, but there were clear differences depending on whether consistency was calculated on the basis of terms or concepts or aspects. Discusses the indexing environment's connections to interindexer consistency

0.010058938 = product of:
  0.030176813 = sum of:
    0.030176813 = weight(_text_:on in 578) [ClassicSimilarity], result of:
      0.030176813 = score(doc=578,freq=4.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.27492687 = fieldWeight in 578, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0625 = fieldNorm(doc=578)
  0.33333334 = coord(1/3)

Source

From classification to 'knowledge organization': Dorking revisited or 'past is prelude'. A collection of reprints to commemorate the firty year span between the Dorking Conference (First International Study Conference on Classification Research 1957) and the Sixth International Study Conference on Classification Research (London 1997). Ed.: A. Gilchrist

0.009940362 = product of:
  0.029821085 = sum of:
    0.029821085 = weight(_text_:on in 7151) [ClassicSimilarity], result of:
      0.029821085 = score(doc=7151,freq=10.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.271686 = fieldWeight in 7151, product of:
          3.1622777 = tf(freq=10.0), with freq of:
            10.0 = termFreq=10.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=7151)
  0.33333334 = coord(1/3)

Abstract

This article revisits the debate concerning controlled and natural indexing languages, as used in searching the databases of the online hosts, in-house information retrieval systems, online public access catalogues and databases stored on CD-ROM. The debate was first formulated in the early days of information retrieval more than a century ago but, despite significant advance in technology, remains unresolved. The article divides the history of the debate into four eras. Era one was characterised by the introduction of controlled vocabulary. Era two focused on comparisons between different indexing languages in order to assess which was best. Era three saw a number of case studies of limited generalisability and a general recognition that the best search performance can be achieved by the parallel use of the two types of indexing languages. The emphasis in Era four has been on the development of end-user-based systems, including online public access catalogues and databases on CD-ROM. Recent developments in the use of expert systems techniques to support the representation of meaning may lead to systems which offer significant support to the user in end-user searching. In the meantime, however, information retrieval in practice involves a mixture of natural and controlled indexing languages used to search a wide variety of different kinds of databases

0.009239726 = product of:
  0.027719175 = sum of:
    0.027719175 = weight(_text_:on in 1054) [ClassicSimilarity], result of:
      0.027719175 = score(doc=1054,freq=6.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.25253648 = fieldWeight in 1054, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.046875 = fieldNorm(doc=1054)
  0.33333334 = coord(1/3)

Abstract

This article presents the results of research into the automatic selection of Library of Congress Classification numbers based on the titles and subject headings in MARC records. The method used in this study was based on partial match retrieval techniques using various elements of new recors (i.e., those to be classified) as "queries", and a test database of classification clusters generated from previously classified MARC records. Sixty individual methods for automatic classification were tested on a set of 283 new records, using all combinations of four different partial match methods, five query types, and three representations of search terms. The results indicate that if the best method for a particular case can be determined, then up to 86% of the new records may be correctly classified. The single method with the best accuracy was able to select the correct classification for about 46% of the new records.

0.009239726 = product of:
  0.027719175 = sum of:
    0.027719175 = weight(_text_:on in 5262) [ClassicSimilarity], result of:
      0.027719175 = score(doc=5262,freq=6.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.25253648 = fieldWeight in 5262, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.046875 = fieldNorm(doc=5262)
  0.33333334 = coord(1/3)

Abstract

The study was based on the collection examnined by W.H. Shaw (Inf. proc. man. 26(1990) no.6, S.693-703, 705-718), a test collection of 1239 articles, indexed with the term cystic fibrosis; and 100 queries with 3 sets of relevance evaluations from subject experts. The effect of variations in indexing exhaustivity on retrieval performance in a vector space retrieval system was investigated by using a term weight threshold to construct different document representations for a test collection. Retrieval results showed that retrieval performance, as measured by the mean optimal measure for all queries at a term weight threshold, was highest at the most exhaustive representation, and decreased slightly as terms were eliminated and the indexing representation became less exhaustive. The findings suggest that the vector space model is more robust against variations in indexing exhaustivity that is the single-link clustering model

0.009015355 = product of:
  0.027046064 = sum of:
    0.027046064 = product of:
      0.054092128 = sum of:
        0.054092128 = weight(_text_:22 in 7247) [ClassicSimilarity], result of:
          0.054092128 = score(doc=7247,freq=2.0), product of:
            0.1747608 = queryWeight, product of:
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.04990557 = queryNorm
            0.30952093 = fieldWeight in 7247, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.0625 = fieldNorm(doc=7247)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Source

Canadian journal of information and library science. 21(1996) no.2, S.1-22

0.008801571 = product of:
  0.026404712 = sum of:
    0.026404712 = weight(_text_:on in 5612) [ClassicSimilarity], result of:
      0.026404712 = score(doc=5612,freq=4.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.24056101 = fieldWeight in 5612, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=5612)
  0.33333334 = coord(1/3)

Abstract

Recent papers have called attention to discrepancies in the assignment of LCSH. While philosophical arguments can be made that subject analysis, if not a logical impossibility, at least is point-of-view dependent, subject headings continue to be assigned and continue to be useful. The hypothesis advanced in the present project is that to a considerable degree there is a clear-cut right and wrong to LCSH subject heading assignment. To test the hypothesis, it was postulated that the assignment of a subject heading is correct if it is supported by textual warrant (at least 20% of the book being cataloged is on the topic) and is constructed in accordance with the LoC Subject Cataloging Manual: Subject Headings. A sample of 100 books on scientific subjects was used to test the hypothesis

0.008801571 = product of:
  0.026404712 = sum of:
    0.026404712 = weight(_text_:on in 1830) [ClassicSimilarity], result of:
      0.026404712 = score(doc=1830,freq=4.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.24056101 = fieldWeight in 1830, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=1830)
  0.33333334 = coord(1/3)

Abstract

Automatic keyword extraction is an important and fundamental technology in an advanced information retrieval systems. Briefly compares several major keyword extraction methods, lists their advantages and disadvantages, and reports recent research progress in Taiwan. Also describes the application of a keyword extraction algorithm in an information retrieval system for relevance feedback. Preliminary analysis shows that the error rate of extracting relevant keywords is 18%, and that the precision rate is over 50%. The main disadvantage of this approach is that the extraction results depend on the retrieval results, which in turn depend on the data held by the database. Apart from collecting more data, this problem can be alleviated by the application of a thesaurus constructed by the same keyword extraction algorithm

0.007888435 = product of:
  0.023665305 = sum of:
    0.023665305 = product of:
      0.04733061 = sum of:
        0.04733061 = weight(_text_:22 in 3510) [ClassicSimilarity], result of:
          0.04733061 = score(doc=3510,freq=2.0), product of:
            0.1747608 = queryWeight, product of:
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.04990557 = queryNorm
            0.2708308 = fieldWeight in 3510, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.0546875 = fieldNorm(doc=3510)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Source

Health libraries review. 7(1990) no.1, S.22-26

0.0076997704 = product of:
  0.02309931 = sum of:
    0.02309931 = weight(_text_:on in 1066) [ClassicSimilarity], result of:
      0.02309931 = score(doc=1066,freq=6.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.21044704 = fieldWeight in 1066, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=1066)
  0.33333334 = coord(1/3)

Abstract

Reports 3 explorations of ranked system design. 2 tests used a 'cystic fibrosis' test collection with 100 queries. Experiment 1 compared a Boolean with a ranked interactive system using a subject qualified trained searcher, and reporting recall and precision results. Experiment 2 compared 15 different ranked match algorithms in a batch mode using 2 test collections, and included some new proximate pairs and term weighting approaches. Experiment 3 is a design plan for an interactive ranked prototype offering mid search algorithm choices plus other manual search devices (such as obligatory and unwanted terms), as influenced by thinking aloud comments from experiment 1. Concludes that, in Boolean versus ranked using inverse collection frequency, the searcher inspected more records on ranked than Boolean and so achieved a higher recall but lower precision; however, the presentation order of the relevant records, was, on average, very similar in both systems. Concludes also that: query reformulation was quite strongly practised in ranked searching but does not appear to have been effective; the term pairs proximate weithing methods in experiment 2 enhanced precision on both test collections when used with inverse collection frequency weighting (ICF); and the design plan for an interactive prototype adds to a selection of match algorithms other devices, such as obligatory and unwanted term marking, evidence for this being found from think aloud comments

0.0075442037 = product of:
  0.02263261 = sum of:
    0.02263261 = weight(_text_:on in 2638) [ClassicSimilarity], result of:
      0.02263261 = score(doc=2638,freq=4.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.20619515 = fieldWeight in 2638, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.046875 = fieldNorm(doc=2638)
  0.33333334 = coord(1/3)

Abstract

Studies the quality of subject indexing by Chemical Abstracts Indexing Service by confronting authors with the particular indexing terms attributed to their computer, for 270 articles published in 54 journals, 5 articles out of each journal. Responses (80%) indicate the superior quality of keywords, both as content descriptors and as retrieval tools. Author judgements on these 2 different aspects do not always converge, however. CAS's indexing policy to cover only 'new' aspects is reflected in author's judgements that index lists are somewhat incomplete, in particular in the case of thesaurus terms (index headings). The large effort expanded by CAS in maintaining and using a subject thesuaurs, in order to select valid index headings, as compared to quick and cheap keyword postings, does not lead to clear superior quality of thesaurus terms for document description nor in retrieval. Some 20% of papers were not placed in 'proper' CA main section, according to authors. As concerns the use of indexing data by third parties, in bibliometrics, users should be aware of the indexing policies behind the data, in order to prevent invalid interpretations

0.007112743 = product of:
  0.021338228 = sum of:
    0.021338228 = weight(_text_:on in 4778) [ClassicSimilarity], result of:
      0.021338228 = score(doc=4778,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.19440265 = fieldWeight in 4778, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0625 = fieldNorm(doc=4778)
  0.33333334 = coord(1/3)

Abstract

Online public access catalogues enable users to undertake subject searching by classification schedules, natural language, or controlled language terminology. In practice the 1st method is little used. Controlled language systems require indexers to index specifically and consistently. A comparative survey was made of indexing practices at Amsterdam and Mijmegen university libraries. On average Amsterdam assigned each document 3.5 index terms against 1.8 at Nijmegen. This discrepancy in indexing policy is the result of long-standing practices in each institution. Nijmegen has failed to utilise the advantages offered by online cataloges

0.007112743 = product of:
  0.021338228 = sum of:
    0.021338228 = weight(_text_:on in 3833) [ClassicSimilarity], result of:
      0.021338228 = score(doc=3833,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.19440265 = fieldWeight in 3833, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0625 = fieldNorm(doc=3833)
  0.33333334 = coord(1/3)

Abstract

Presents results of an experiment in which 8 indexers (4 beginners and 4 experts) were asked to index the same 4 documents with 2 different thesauri. The 3 kind of verbal reports provide complementary data on strategic behaviour. it is of prime importance to consider the indexing task as an ill-defined problem, where the solutionm is partly defined by the indexer

0.0062868367 = product of:
  0.01886051 = sum of:
    0.01886051 = weight(_text_:on in 4152) [ClassicSimilarity], result of:
      0.01886051 = score(doc=4152,freq=4.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.1718293 = fieldWeight in 4152, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=4152)
  0.33333334 = coord(1/3)

Abstract

A software system has been developed which orders citations retrieved from an online database in terms of relevancy. The system resulted from an effort generated by NASA's Technology Utilization Program to create new advanced software tools to largely automate the process of determining relevancy of database citations retrieved to support large technology transfer studies. The ranking is based on the generation of an enriched vocabulary using lexical association methods, a user assessment of the vocabulary and a combination of the user assessment and the lexical metric. One of the key elements in relevancy ranking is the enriched vocabulary -the terms mst be both unique and descriptive. This paper examines term uniqueness. Six lexical association methods were employed to generate characteristic word indices. A limited subset of the terms - the highest 20,40,60 and 7,5% of the uniquess words - we compared and uniquess factors developed. Computational times were also measured. It was found that methods based on occurrences and signal produced virtually the same terms. The limited subset of terms producedby the exact and centroid discrimination value were also nearly identical. Unique terms sets were produced by teh occurrence, variance and discrimination value (centroid), An end-user evaluation showed that the generated terms were largely distinct and had values of word precision which were consistent with values of the search precision.

0.0062868367 = product of:
  0.01886051 = sum of:
    0.01886051 = weight(_text_:on in 554) [ClassicSimilarity], result of:
      0.01886051 = score(doc=554,freq=4.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.1718293 = fieldWeight in 554, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0390625 = fieldNorm(doc=554)
  0.33333334 = coord(1/3)

Abstract

The development of regional on-line catalogs and in-house information systems for retrieval of references provide examples of the impact of indexing theory and applications on technical processing. More emphasis must be given to understanding the techniques for evaluating the effectiveness of a file, irrespective of whether that file was created as a library catalog or an index to information sources. The most significant advances in classification theory in recent decades has been as a result of efforts to improve effectiveness of indexing systems. Library classification systems are indexing languages or systems. Courses offered for the preparation of indexers in the United States and the United Kingdom are reviewed. A point of congruence for both the indexer and the library classifier would appear to be the need for a thorough preparation in the techniques of subject analysis. Any subject heading list will suffer from omissions as well as the inclusion of terms which the patron will never use. Indexing theory has provided the technical services department with methods for evaluation of effectiveness. The writer does not believe that these techniques are used, nor do current courses, workshops, and continuing education programs stress them. When theory is totally subjugated to practice, critical thinking and maximum effectiveness will suffer.

0.00622365 = product of:
  0.01867095 = sum of:
    0.01867095 = weight(_text_:on in 4526) [ClassicSimilarity], result of:
      0.01867095 = score(doc=4526,freq=2.0), product of:
        0.109763056 = queryWeight, product of:
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.04990557 = queryNorm
        0.17010231 = fieldWeight in 4526, product of:
          1.4142135 = tf(freq=2.0), with freq of:
            2.0 = termFreq=2.0
          2.199415 = idf(docFreq=13325, maxDocs=44218)
          0.0546875 = fieldNorm(doc=4526)
  0.33333334 = coord(1/3)

Abstract

Reports results of a study of 2 information retrieval systems on a 2.000 document full text medical database. The first system, SAPHIRE, features concept based automatic indexing and statistical retrieval techniques, while the second system, SWORD, features traditional word based Boolean techniques, 16 medical students at Oregon Health Sciences Univ. each performed 10 searches and their results, recorded in terms of recall and precision, showed nearly equal performance for both systems. SAPHIRE was also compared with a version of SWORD modified to use automatic indexing and ranked retrieval. Using batch input of queries, the latter method performed slightly better