IRS13 Search Matching Functions chapter E. M. Keen Harvard University Gerard Salton Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. 111-53 NUMBER AND PERCENTAGE EVALUATION OF INDIVIDUAL REQUESTS COLLECTION INPUT AND DICTIONARY MEASURE USED COSINE COSINE LOGICAL, TO DETERMINE NUMERIC OVERLAP LOGICAL MERIT SUEERIOR OR OVERLAP NUM- ERIC SUPERIOR IPE-3 Abstract, Stem Normed Recall 19 (55.9%) 15 (144.i%) 3[OCRerr] Normed Preci- Requests sion 23 (67.6%) 11 (32.[OCRerr]) CRAN-1 Abstract, Stem Normed Recall 33 (78.6% 9 (21.4%) [OCRerr]2 Normed Preci- Requests sion 28 (66.7%) l[OCRerr] (33.3%) [OCRerr] Comparisons of individual request merit giving the nun[OCRerr]bers of requests favoring cosine numeric versus those favoring any of the other three functions, on two collections, according to merit assigned by normalized recall and precision. Figure 3[OCRerr].