Sager, W.K.H.; Lockemann, P.C.: Classification of ranking algorithms (1976)
0.02
0.024480894 = product of:
0.14688537 = sum of:
0.14688537 = weight(_text_:ranking in 7404) [ClassicSimilarity], result of:
0.14688537 = score(doc=7404,freq=6.0), product of:
0.20271951 = queryWeight, product of:
5.4090285 = idf(docFreq=537, maxDocs=44218)
0.03747799 = queryNorm
0.7245744 = fieldWeight in 7404, product of:
2.4494898 = tf(freq=6.0), with freq of:
6.0 = termFreq=6.0
5.4090285 = idf(docFreq=537, maxDocs=44218)
0.0546875 = fieldNorm(doc=7404)
0.16666667 = coord(1/6)
- Abstract
- The retireval of documents by content is based on identifiers that are seldom unique to a document. Therefore, a query is usually answered by a number of documents that may, however, differ in relevance to the query. Ranking algorithms attempt to assess these relevances by calculating a relevance measure, and order the documents on the basis of the measure. Surveys the algorithms found in the literature for the weighting of index terms and the calculation of relevance measures, classifies them according to various criteria, and compares a number of them empirically. Methods are discussed for further classification of the algorithms on the basis of similarity of behaviour. These are demonstrated for a data base of 1.003 legal documents. It is shown that one may considerably improve the quality of ranking by combining algorithms from different classes
