Search (22 results, page 1 of 2)

0.010863388 = product of:
  0.032590162 = sum of:
    0.032590162 = weight(_text_:to in 5545) [ClassicSimilarity], result of:
      0.032590162 = score(doc=5545,freq=12.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.39361957 = fieldWeight in 5545, product of:
          3.4641016 = tf(freq=12.0), with freq of:
            12.0 = termFreq=12.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0625 = fieldNorm(doc=5545)
  0.33333334 = coord(1/3)

Abstract

Answers to questions concerning individual library decisions to adopt classification systems are important in understanding the efffectiveness of libraries but are difficult to provide. Measures of classification system performance are discussed, as are different methodologies that may be used to seek answers, ranging from formal or philosophical models to quantitative experimental techniques and qualitative methods

0.009916872 = product of:
  0.029750613 = sum of:
    0.029750613 = weight(_text_:to in 4508) [ClassicSimilarity], result of:
      0.029750613 = score(doc=4508,freq=10.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.3593239 = fieldWeight in 4508, product of:
          3.1622777 = tf(freq=10.0), with freq of:
            10.0 = termFreq=10.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0625 = fieldNorm(doc=4508)
  0.33333334 = coord(1/3)

Abstract

For classification to advance to the point where optimal systems may be developed for manual or automated use, it will be necessary for a science of document or library classification to be developed. Seven questions are posed which the author feels must be answered before such optimal systems can be developed. Suggestions are made as to the forms that answers to these questions might take

0.00940797 = product of:
  0.02822391 = sum of:
    0.02822391 = weight(_text_:to in 146) [ClassicSimilarity], result of:
      0.02822391 = score(doc=146,freq=16.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.34088457 = fieldWeight in 146, product of:
          4.0 = tf(freq=16.0), with freq of:
            16.0 = termFreq=16.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=146)
  0.33333334 = coord(1/3)

Abstract

Relevance and economic feedback may be used to produce an ordering of documents that supports browsing in hypermedia and digital libraries. Document classification based on the Gray code provides paths through the entire collection, each path traversing each node in the set of documents exactly once. Examines systems organizing document based on weighted and unweighted Gray codes. Relevance feedback is used to conceptually organize the collection for an individual to browse, based on that individual's interests and information needs, as reflected by their relevance judgements and user supplied economic preferences. Applies Bayesian learning theory to estimating the characteristics of documents of interest to the user and supplying an analytic model of browsing performance, based on minimising the Expected Browsing Distance. Economic feedback may be used to change the ordering of documents to benefit the user. Using these techniques, a hypermedia or digital library may order any and all available documents, not just those examined, based on the information provided by the searcher or people with similar interests

Footnote

Contribution to a special issue on methods and tools for the automatic construction of hypertext

0.0076815756 = product of:
  0.023044726 = sum of:
    0.023044726 = weight(_text_:to in 7441) [ClassicSimilarity], result of:
      0.023044726 = score(doc=7441,freq=6.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.2783311 = fieldWeight in 7441, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0625 = fieldNorm(doc=7441)
  0.33333334 = coord(1/3)

Abstract

State of the art review of the mechanisms of feedback in information retrieval (IR) in terms of feedback concepts and models in cybernetics and social sciences. Critically evaluates feedback research based on the traditional IR models and comparing the different approaches to automatic relevance feedback techniques, and feedback research within the framework of interactive IR models. Calls for an extension of the concept of feedback beyond relevance feedback to interactive feedback. Cites specific examples of feedback models used within IR research and presents 6 challenges to future research

0.0074376534 = product of:
  0.02231296 = sum of:
    0.02231296 = weight(_text_:to in 8525) [ClassicSimilarity], result of:
      0.02231296 = score(doc=8525,freq=10.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.26949292 = fieldWeight in 8525, product of:
          3.1622777 = tf(freq=10.0), with freq of:
            10.0 = termFreq=10.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=8525)
  0.33333334 = coord(1/3)

Abstract

A text window is a group of words appearing in contiguous positions in text used to exploit a variety of lexical, syntactics, and semantic relationships without having to analyze the text explicitely for their structure. This supports the previously suggested idea that natural grouping of words are best treated as a unit of size 7 to 11 words, that is, plus or minus 3 to 5 words. The text retrieval experiments varying the size of windows, both with full text and with stopwords removed, support these size ranges. The characteristcs of windows that best match terms in queries are examined in detail, revealing intersting differences between those for queries with good results and those for queries with poorer results. Queries with good results tend to contain morte content word phrase and few terms with high frequency of use in the database. Information retrieval systems may benefit from expanding thesaurus-style relationships or incorporating statistical dependencies for terms within these windows

0.0074376534 = product of:
  0.02231296 = sum of:
    0.02231296 = weight(_text_:to in 4162) [ClassicSimilarity], result of:
      0.02231296 = score(doc=4162,freq=10.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.26949292 = fieldWeight in 4162, product of:
          3.1622777 = tf(freq=10.0), with freq of:
            10.0 = termFreq=10.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=4162)
  0.33333334 = coord(1/3)

Abstract

An analytic method of information retrieval and filtering evaluation can quantitatively predict the expected number of documents examined in retrieving a relevant document. It also allows researchers and practioners to qualitatively understand how varying different estimates of query parameter values affects retrieval performance. The incoorporation of relevance feedback to increase our knowledge about the parameters of relevant documents and the robustness of parameter estimates is modeled. Single term and two term independence models, as well as a complete term dependence model, are developed. An economic model of retrieval performance may be used to study the effects of database size and to provide analytic answers to questions comparing retrieval from small and large databases, as well as questions about the number of terms in a query. Results are presented as a performance surface, a three dimensional graph showing the effects of two independent variables on performance.

0.007198559 = product of:
  0.021595677 = sum of:
    0.021595677 = product of:
      0.043191355 = sum of:
        0.043191355 = weight(_text_:22 in 3368) [ClassicSimilarity], result of:
          0.043191355 = score(doc=3368,freq=2.0), product of:
            0.15947726 = queryWeight, product of:
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.045541126 = queryNorm
            0.2708308 = fieldWeight in 3368, 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=3368)
      0.5 = coord(1/2)
  0.33333334 = coord(1/3)

Date

22. 2.1996 13:14:10

0.006721379 = product of:
  0.020164136 = sum of:
    0.020164136 = weight(_text_:to in 4485) [ClassicSimilarity], result of:
      0.020164136 = score(doc=4485,freq=6.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.24353972 = fieldWeight in 4485, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0546875 = fieldNorm(doc=4485)
  0.33333334 = coord(1/3)

Abstract

Patrons often browse through books organized by a library classification system, looking for books to use and possibly circulate. This research is an examination of the clustering of similar books provided by a classification system and ways in which the books that patrons circulate are clustered. Measures of classification system performance are suggested and used to evaluate two test collections. Regression formulas are derived describing the relationships among the number of areas in which books were found (the number of stops a patron makes when browsing), the distances across a cluster, and the average number of books a patron circulates. Patrons were found usually to make more stops than there were books found at their average stop. Consequences for full-text document systems and online catalogs are suggested

0.006721379 = product of:
  0.020164136 = sum of:
    0.020164136 = weight(_text_:to in 6962) [ClassicSimilarity], result of:
      0.020164136 = score(doc=6962,freq=6.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.24353972 = fieldWeight in 6962, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0546875 = fieldNorm(doc=6962)
  0.33333334 = coord(1/3)

Abstract

Knowledge of window style, content, location, and grammatical structure may be used to classify documents as originating within a particular discipline or may be used to place a document on a theory vs. practice spectrum. Examines characteristics of phrases and text windows, including their number, location in documents, and grammatical construction, in addition to studying variations in these window characteristics across disciplines. Examines some of the linguistic regularities for individual disciplines, and suggests families of regularities that may provide helpful for the automatic classification of documents, as well as for information retrieval and filtering applications

0.006721379 = product of:
  0.020164136 = sum of:
    0.020164136 = weight(_text_:to in 4860) [ClassicSimilarity], result of:
      0.020164136 = score(doc=4860,freq=6.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.24353972 = fieldWeight in 4860, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0546875 = fieldNorm(doc=4860)
  0.33333334 = coord(1/3)

Abstract

The variety of performance measures available for information retrieval systems, search engines, and network filtering agents can be confusing to both practitioners and scholars. Most discussions about these measures address their theoretical foundations and the characteristics of a measure that make it desirable for a particular application. In this work, we consider how measures of performance at a point in a search may be formally compared. Criteria are developed that allow one to determine the percent of time or conditions under which 2 different performance measures suggest that one document ordering is superior to another ordering, or when the 2 measures disagree about the relative value of document orderings. As an example, graphs provide illustrations of the relationships between precision and F

0.00627198 = product of:
  0.018815938 = sum of:
    0.018815938 = weight(_text_:to in 380) [ClassicSimilarity], result of:
      0.018815938 = score(doc=380,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.22725637 = fieldWeight in 380, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0625 = fieldNorm(doc=380)
  0.33333334 = coord(1/3)

Abstract

Information may be defined as the characteristics of the output of a process, these being informative about the process and the input. This discipline independent definition may be applied to all domains, from physics to epistemology. Hierarchies of processes linked together, provide a communication channel between each of the corresponding functions and layers in the hierarchies. Models of communication, perception, observation, belief, and knowledge are suggested that are consistent with this conceptual framework of information as the value of the output of any process in a hierarchy of processes. Misinformation and errors are considered

0.005761182 = product of:
  0.017283546 = sum of:
    0.017283546 = weight(_text_:to in 924) [ClassicSimilarity], result of:
      0.017283546 = score(doc=924,freq=6.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.20874833 = fieldWeight in 924, product of:
          2.4494898 = tf(freq=6.0), with freq of:
            6.0 = termFreq=6.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=924)
  0.33333334 = coord(1/3)

Abstract

A thesaurus and an ontology provide a set of structured terms, phrases, and metadata, often in a hierarchical arrangement, that may be used to index, search, and mine documents. We describe the decisions that should be made when including a term, deciding whether a term should be subdivided into its subclasses, or determining which of more than one set of possible subclasses should be used. Based on retrospective measurements or estimates of future performance when using thesaurus terms in document ordering, decisions are made so as to maximize performance. These decisions may be used in the automatic construction of a thesaurus. The evaluation of an existing thesaurus is described, consistent with the decision criteria developed here. These kinds of user-focused decision-theoretic techniques may be applied to other hierarchical applications, such as faceted classification systems used in information architecture or the use of hierarchical terms in "breadcrumb navigation".

0.0055436995 = product of:
  0.016631098 = sum of:
    0.016631098 = weight(_text_:to in 5148) [ClassicSimilarity], result of:
      0.016631098 = score(doc=5148,freq=8.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.20086816 = fieldWeight in 5148, product of:
          2.828427 = tf(freq=8.0), with freq of:
            8.0 = termFreq=8.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0390625 = fieldNorm(doc=5148)
  0.33333334 = coord(1/3)

Abstract

Documents in digital and paper libraries may be arranged, based on their topics, in order to facilitate browsing. It may seem intuitively obvious that ordering documents by their subject should improve browsing performance; the results presented in this article suggest that ordering library materials by their Gray code values and through using links consistent with the small world model of document relationships is consistent with improving browsing performance. Below, library circulation data, including ordering with Library of Congress Classification numbers and Library of Congress Subject Headings, are used to provide information useful in generating user-centered document arrangements, as well as user-independent arrangements. Documents may be linearly arranged so they can be placed in a line by topic, such as on a library shelf, or in a list on a computer display. Crossover links, jumps between a document and another document to which it is not adjacent, can be used in library databases to allow additional paths that one might take when browsing. The improvement that is obtained with different combinations of document orderings and different crossovers is examined and applications suggested.

0.005487982 = product of:
  0.016463947 = sum of:
    0.016463947 = weight(_text_:to in 2335) [ClassicSimilarity], result of:
      0.016463947 = score(doc=2335,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.19884932 = fieldWeight in 2335, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.0546875 = fieldNorm(doc=2335)
  0.33333334 = coord(1/3)

Abstract

A document classifier places documents together in a linear arrangement for browsing or high-speed access by human or computerised information retrieval systems. Requirements for document classification and browsing systems are developed from similarity measures, distance measures, and the notion of subject aboutness. A requirement that documents be arranged in decreasing order of similarity as the distance from a given document increases can often not be met. Based on these requirements, information-theoretic considerations, and the Gray code, a classification system is proposed that can classifiy documents without human intervention. A measure of classifier performance is developed, and used to evaluate experimental results comparing the distance between subject headings assigned to documents given classifications from the proposed system and the Library of Congress Classification (LCC) system

0.005431694 = product of:
  0.016295081 = sum of:
    0.016295081 = weight(_text_:to in 1016) [ClassicSimilarity], result of:
      0.016295081 = score(doc=1016,freq=12.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.19680978 = fieldWeight in 1016, product of:
          3.4641016 = tf(freq=12.0), with freq of:
            12.0 = termFreq=12.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.03125 = fieldNorm(doc=1016)
  0.33333334 = coord(1/3)

Abstract

Relationships between terms and features are an essential component of thesauri, ontologies, and a range of controlled vocabularies. In this article, we describe ways to identify important concepts in documents using the relationships in a thesaurus or other vocabulary structures. We introduce a methodology for the analysis and modeling of the indexing process based on a weighted random walk algorithm. The primary goal of this research is the analysis of the contribution of thesaurus structure to the indexing process. The resulting models are evaluated in the context of automatic subject indexing using four collections of documents pre-indexed with 4 different thesauri (AGROVOC [UN Food and Agriculture Organization], high-energy physics taxonomy [HEP], National Agricultural Library Thesaurus [NALT], and medical subject headings [MeSH]). We also introduce a thesaurus-centric matching algorithm intended to improve the quality of candidate concepts. In all cases, the weighted random walk improves automatic indexing performance over matching alone with an increase in average precision (AP) of 9% for HEP, 11% for MeSH, 35% for NALT, and 37% for AGROVOC. The results of the analysis support our hypothesis that subject indexing is in part a browsing process, and that using the vocabulary and its structure in a thesaurus contributes to the indexing process. The amount that the vocabulary structure contributes was found to differ among the 4 thesauri, possibly due to the vocabulary used in the corresponding thesauri and the structural relationships between the terms. Each of the thesauri and the manual indexing associated with it is characterized using the methods developed here.

0.004703985 = product of:
  0.014111955 = sum of:
    0.014111955 = weight(_text_:to in 7418) [ClassicSimilarity], result of:
      0.014111955 = score(doc=7418,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.17044228 = fieldWeight in 7418, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=7418)
  0.33333334 = coord(1/3)

Abstract

The best-case, random and worst-case document rankings and retrieval performance may be determined using a method discussed here. Knowledge of the best case performance allows users and system designers to determine how close to the optimum condition their search is and select queries and matching functions that will produce the best results. Suggests a method for deriving the optimal Boolean query for a given level of recall and a method for determining the quality of a Boolean query. Measures are proposed that modify conventional text retrieval measures such as precision, E, and average search length, so that the values for these measures are 1 when retrieval is optimal, 0 when retrieval is random, and -1 when worst-case. Tests using one of these measures show that many retrieval are optimal? Consequences for retrieval research are examined

0.004703985 = product of:
  0.014111955 = sum of:
    0.014111955 = weight(_text_:to in 4068) [ClassicSimilarity], result of:
      0.014111955 = score(doc=4068,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.17044228 = fieldWeight in 4068, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=4068)
  0.33333334 = coord(1/3)

Abstract

The grammars of natural languages may be learned by using genetic algorithms that reproduce and mutate grammatical rules and parts of speech tags, improving the quality of later generations of grammatical components. Syntactic rules are randomly generated and then evolve; those rules resulting in improved parsing and occasionally improved filtering performance are allowed to further propagate. The LUST system learns the characteristics of the language or subkanguage used in document abstracts by learning from the document rankings obtained from the parsed abstracts. Unlike the application of traditional linguistic rules to retrieval and filtering applications, LUST develops grammatical structures and tags without the prior imposition of some common grammatical assumptions (e.g. part of speech assumptions), producing grammars that are empirically based and are optimized for this particular application

0.004703985 = product of:
  0.014111955 = sum of:
    0.014111955 = weight(_text_:to in 6976) [ClassicSimilarity], result of:
      0.014111955 = score(doc=6976,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.17044228 = fieldWeight in 6976, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=6976)
  0.33333334 = coord(1/3)

Abstract

There are regularities in the statistical information provided by natural language terms about neighboring terms. We find that when phrase rank increases, moving from common to less common phrases, the value of the expected mutual information measure (EMIM) between the terms regularly decreases. Luhn's model suggests that midrange terms are the best index terms and relevance discriminators. We suggest reasons for this principle based on the empirical relationships shown here between the rank of terms within phrases and the average mutual information between terms, which we refer to as the Inverse Representation- EMIM principle. We also suggest an Inverse EMIM term weight for indexing or retrieval applications that is consistent with Luhn's distribution. An information theoretic interpretation of Zipf's Law is provided. Using the regularity noted here, we suggest that Zipf's Law is a consequence of the statistical dependencies that exist between terms, described here using information theoretic concepts.

0.004703985 = product of:
  0.014111955 = sum of:
    0.014111955 = weight(_text_:to in 173) [ClassicSimilarity], result of:
      0.014111955 = score(doc=173,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.17044228 = fieldWeight in 173, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=173)
  0.33333334 = coord(1/3)

Abstract

Both structured and unstructured data, as well as structured data representing several different types of tuples, may be integrated into a single list for browsing or retrieval. Data may be arranged in the Gray code order of the features and metadata, producing optimal ordering for browsing. We provide several metrics for evaluating the performance of systems supporting browsing, given some constraints. Metadata and indexing terms are used for sorting keys and attributes for structured data, as well as for semi-structured or unstructured documents, images, media, etc. Economic and information theoretic models are suggested that enable the ordering to adapt to user preferences. Different relational structures and unstructured data may be integrated into a single, optimal ordering for browsing or for displaying tables in digital libraries, database management systems, or information retrieval systems. Adaptive displays of data are discussed.

0.004703985 = product of:
  0.014111955 = sum of:
    0.014111955 = weight(_text_:to in 1100) [ClassicSimilarity], result of:
      0.014111955 = score(doc=1100,freq=4.0), product of:
        0.08279609 = queryWeight, product of:
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.045541126 = queryNorm
        0.17044228 = fieldWeight in 1100, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.818051 = idf(docFreq=19512, maxDocs=44218)
          0.046875 = fieldNorm(doc=1100)
  0.33333334 = coord(1/3)

Abstract

The assignment of indexing terms and metadata to documents, data, and other information representations is considered useful, but the utility of including a single term is seldom discussed. The author discusses a simple model of document ordering and then shows how assigning index and metadata labels improves or decreases retrieval performance. The Indexing and Metadata Advantage (IMA) factor measures how indexing or assigning a metadata term helps (or hurts) ordering performance. Performance values and the associated IMA expressions are computed, consistent with several different assumptions. The economic value associated with various term assignment decisions is developed. The IMA term advantage model itself is empirically validated with computer software that shows that the analytic results obtained agree completely with the actual performance gains and losses found when ordering all sets of 14 or fewer documents. When the formulas in the software are changed to differ from this model, the predictions of the actual performance are erroneous.