IRS13 Evaluation Parameters 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. -- [OCRerr]-- Suffix £5' Thesaurus[OCRerr]2 (New Q. S.) iuuIg::uI Be St possible performance for Cumulated any dictionary Relevance Points 44[OCRerr]02[OCRerr] 360 320 280 240 200 160 120 80 40 I I I I I I I I I I I [OCRerr] Rank I 3 5 10 1520305075 150 Positions 100 200 [OCRerr][OCRerr]Suffix `S' [OCRerr]>Thesaurus-2 (New Q.S.) Best possible performance fo any dictionary Precision 1.0 - 0.8 0.6 0.4 0.2 O~---O~O ¼ - I 3 - O[OCRerr]q[OCRerr] [OCRerr] I I[OCRerr]I3035[OCRerr] 2[OCRerr]0 Re I.0 0 0.2 0.4 0.6 0.8 Cran-l, Abstracts, Micro Evaluation over 42 Requests. a) Performance characteristic curve using relevance points b) Precision recall curve, with recall on relevance points, "Pseudo-Cranf cut-off. Performance Curves that are [OCRerr]ased on Grades of Relevance Fig. 26