IRS13 Thesaurus, Phrase and Hierarchy Dictionaries 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. VII-23 Number and Percentage Evaluation, of Individual Requests Based on Collection, Input Average Rank, Thesaurus Stem Both and Thesaurus Type Fig. 14 Superior Superio[OCRerr] Equal IRE-3, Abstract First Rel. 5 14.7% 5 14.7% 24 70.6% Second Rel. 11 32.4% 4 11.8% 19 55.9% (Thesaurus-3), 34 Requests Last Rel. 27 79.4% 7 20.6% 0 - CRAN-1, Abstract First Rel. 12 28.6% 9 21.4% 11 26.2% Second Rel.* 14 34.1% 12 29.3% 15 36.6% (Thesaurus-3), 42 Requests Last Rel. 24 57.1% 15 35.7% 7 16.7% First Rel. 15 35.7% 7 16.7% 20 47.6% CRAN-l, Indexing, Second Rel.* 21 51.2% 7 17.1% 13 31.7% (Thesaurus-2), 42 Requests LastRel. 18 42.9% 21 50.0% 3 7.1% First Rel. 17 48.6% 6 17.1% 12 34.3% ADI, Text, Second Rel.t 16 51.6% 8 25.8% 7 22.6% (Thesaurus-l), 35 Requests Last Rel. 20 57.1% 13 37.1% 2 5.7% First Rel. 19 54.3% 7 20.0% 9 25.7% ADI, Abstract, Second Rel.i 16 51.6% 11 35.5% 4 12.9% (Thesaurus-i), 35 Requests Last Rel. 20 57.1% 13 37.1% 2 5.7% * In the Cran-l Collection, 1 request has no second relevant, so results are based on 41 requests. In the ADI Collection, 4 requests have no second relevant, so results are based on 31 requests. Comparison of Individual Request Merit Giving the Numbers of Requests Favoring Stem and Thesaurus on Five Runs in Three Collections, according to Merit Assigned by the Average Ranks given in Figure 14 Fig. 15