Search (6 results, page 1 of 1)

0.023891509 = product of:
  0.071674526 = sum of:
    0.071674526 = product of:
      0.10751179 = sum of:
        0.07294742 = weight(_text_:retrieval in 3365) [ClassicSimilarity], result of:
          0.07294742 = score(doc=3365,freq=16.0), product of:
            0.15433937 = queryWeight, product of:
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.051022716 = queryNorm
            0.47264296 = fieldWeight in 3365, product of:
              4.0 = tf(freq=16.0), with freq of:
                16.0 = termFreq=16.0
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.0390625 = fieldNorm(doc=3365)
        0.03456437 = weight(_text_:22 in 3365) [ClassicSimilarity], result of:
          0.03456437 = score(doc=3365,freq=2.0), product of:
            0.17867287 = queryWeight, product of:
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.051022716 = queryNorm
            0.19345059 = fieldWeight in 3365, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              3.5018296 = idf(docFreq=3622, maxDocs=44218)
              0.0390625 = fieldNorm(doc=3365)
      0.6666667 = coord(2/3)
  0.33333334 = coord(1/3)

Abstract

The retrieval effectiveness of 5 hierarchical clustering methods (single link, complete link, group average, Ward's method, and weighted average) is examined as a function of indexing exhaustivity with 4 test collections (CR, Cranfield, Medlars, and Time). Evaluations of retrieval effectiveness, based on 3 measures of optimal retrieval performance, confirm earlier findings that the performance of a retrieval system based on single link clustering varies as a function of indexing exhaustivity but fail ti find similar patterns for other clustering methods. The data also confirm earlier findings regarding the poor performance of single link clustering is a retrieval environment. The poor performance of single link clustering appears to derive from that method's tendency to produce a small number of large, ill defined document clusters. By contrast, the data examined here found the retrieval performance of the other clustering methods to be general comparable. The data presented also provides an opportunity to examine the theoretical limits of cluster based retrieval and to compare these theoretical limits to the effectiveness of operational implementations. Performance standards of the 4 document collections examined were found to vary widely, and the effectiveness of operational implementations were found to be in the range defined as unacceptable. Further improvements in search strategies and document representations warrant investigations

Date

22. 2.1996 11:20:06

0.01123099 = product of:
  0.03369297 = sum of:
    0.03369297 = product of:
      0.10107891 = sum of:
        0.10107891 = weight(_text_:retrieval in 2946) [ClassicSimilarity], result of:
          0.10107891 = score(doc=2946,freq=12.0), product of:
            0.15433937 = queryWeight, product of:
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.051022716 = queryNorm
            0.6549133 = fieldWeight in 2946, product of:
              3.4641016 = tf(freq=12.0), with freq of:
                12.0 = termFreq=12.0
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.0625 = fieldNorm(doc=2946)
      0.33333334 = coord(1/3)
  0.33333334 = coord(1/3)

Abstract

The ability to distinguish between acceptable and unacceptable levels of retrieval performance and the ability to distinguish between significant and non-significant differences between retrieval result are important to traditional information retrieval experiments. The Monte Carlo method is shown to represent an attractive alternative to the hypergeometric model for identifying the levels at which random retrieval performance is exceeded in retrieval test collections and for overcoming some of the limitations of the hypergeometric model

0.010252445 = product of:
  0.030757334 = sum of:
    0.030757334 = product of:
      0.092272 = sum of:
        0.092272 = weight(_text_:retrieval in 3308) [ClassicSimilarity], result of:
          0.092272 = score(doc=3308,freq=10.0), product of:
            0.15433937 = queryWeight, product of:
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.051022716 = queryNorm
            0.59785134 = fieldWeight in 3308, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.0625 = fieldNorm(doc=3308)
      0.33333334 = coord(1/3)
  0.33333334 = coord(1/3)

Abstract

Relevance judgements used to evaluate the performance of information retrieval systems are known to vary among judges and to vary under certain conditions extraneous to the relevance relationship between queries and documents. Investigates the degree to which variations in relevance judgements affect the evaluation of retrieval performance. Four sets of relevance judgements were used to test the retrieval effectiveness of 6 document representations. In no case was there a noticeable or material difference in retrieval performance due to variations in relevance judgement. Detailed examination of reasons why variations in relevance judgements may not affect recall and precision

0.008970889 = product of:
  0.026912667 = sum of:
    0.026912667 = product of:
      0.080738 = sum of:
        0.080738 = weight(_text_:retrieval in 7256) [ClassicSimilarity], result of:
          0.080738 = score(doc=7256,freq=10.0), product of:
            0.15433937 = queryWeight, product of:
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.051022716 = queryNorm
            0.5231199 = fieldWeight in 7256, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.0546875 = fieldNorm(doc=7256)
      0.33333334 = coord(1/3)
  0.33333334 = coord(1/3)

Abstract

Computes low performance standards for the group of queries in 13 retrieval test collections. Derived from the random graph hypothesis, these standards represent the highest levels of retrieval effectivenss that can be obtained from meaningless clustering structures. Compares operational levels of cluster-based performance reported in selected sources during the past 20 years to the standards. Typical levels of operational cluster-based retrieval can be explained on the basis of change. Most operational results in retrieval test collections are lower than those predicted by random graph theory. Clustering strategies capable of adapting to relevance information may succeed where static clustering techniques have failed

0.007689334 = product of:
  0.023068001 = sum of:
    0.023068001 = product of:
      0.069204 = sum of:
        0.069204 = weight(_text_:retrieval in 7259) [ClassicSimilarity], result of:
          0.069204 = score(doc=7259,freq=10.0), product of:
            0.15433937 = queryWeight, product of:
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.051022716 = queryNorm
            0.44838852 = fieldWeight in 7259, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.046875 = fieldNorm(doc=7259)
      0.33333334 = coord(1/3)
  0.33333334 = coord(1/3)

Abstract

Computes low performance standards for each query and for the group of queries in 13 traditional and 4 TREC test collections. Predicted by the hypergeometric distribution, the standards represent the highest level of retrieval effectiveness attributable to chance. Compares operational levels of performance for vector-space, ad-hoc-feature-based, probabilistic, and other retrieval models to the standards. The effectiveness of these techniques in small, traditional test collections, can be explained by retrieving a few more relevant documents for most queries than expected by chance. The effectiveness of retrieval techniques in the larger TREC test collections can only be explained by retrieving many more relevant documents for most queries than expected by chance. The discrepancy between deviations form chance in traditional and TREC test collections is due to a decrease in performance standards for large test collections, not to an increase in operational performance. The next generation of information retrieval systems would be enhanced by abandoning uninformative performance summaries and focusing on effectiveness and improvements in effectiveness of individual queries

0.007689334 = product of:
  0.023068001 = sum of:
    0.023068001 = product of:
      0.069204 = sum of:
        0.069204 = weight(_text_:retrieval in 5262) [ClassicSimilarity], result of:
          0.069204 = score(doc=5262,freq=10.0), product of:
            0.15433937 = queryWeight, product of:
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.051022716 = queryNorm
            0.44838852 = fieldWeight in 5262, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              3.024915 = idf(docFreq=5836, maxDocs=44218)
              0.046875 = fieldNorm(doc=5262)
      0.33333334 = coord(1/3)
  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