CRANV2 Aslib Cranfield Research Project: Factors Determining the Performance of Indexing Systems: Volume 2 Main test results chapter Cyril Cleverdon Michael Keen Cranfield An investigation supported by a grant to Aslib by the National Science Foundation. Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. - 145 - FIGURE 4.61 09' Index Language i, I. a ExLaustlvity of Indexing 3 Search Ru2e A Document Relevance 1-4 Number of Documents in Collection Number of Questions 50 Number of Re!evant Documents 361 Generality Number 5.2 1400 ' Coord - Lrmtion Level i 2 3 4 D 6 7 8 9 10 11 12 Documents Retrieved Rel. Non-tel. 340 30,778 281 11,572 176 4,717 107 i, 727 59 717 35 273 19 80 7 17 5 2 1 0 0 0 0 0 Recall Ratio a/a+c 94.2% 77.8% 48.8% 29.6% 16.3% 9.7% 5.3% I. 9% 1.4% 0.3% Precision Ratio a/a+b 1.1% 2.4% 3.6% 5.6% 7.0% 11.4% 19.2% 29.2% 71.4% 1 oo. 0% Fallout Ratio b/b+d 44.196% 16.017% 6.759% 2.48o% I. 015% 0. 392% 0,115% 0. 024% 0.003% 0. 000% FIGURE 4.011T Index Language I. 1. a Exhaustivity of Indexing 3 Search Rule A Document Relevance 1-3 Number of Documents in Collection 1400 Number of Questions 50 Number of Relevant Documents 297 GeneraLity Number 4.2 I ; 3 4 5 6 7 8 9 10 11 12 Documents Retrieved Rel. 277 235 i02 97 49 20 "16 6 4 1 0 0 Non-rel. 30,041 93.2% 11,618 79.1% 4,731 54.5% 1,737 32.7% 724 16.5% 279 9.8% 83 5.3% 10 2.0% 3 1.3% 0 0.3% 0 0 Recall Ratio a/a+e Precision Ratio a]a+b o. 0% 2.0% 3.4% 5.3% 6.4% 9.2% 16.2% 26.0% 57.1% 1 00. O% Fallout Ratio b/b+d 44. 246[OCRerr] 16.651% 6. 787% 2.492% 1. 039% O. 400% 0.119% O. 026% o. 004% o. 000% 5O 50 47 39 21 12 8 4 3 1 0 0 50 50 50 50 50 50 47 50 50 39 46 46 21 40 40 12 30 30 8 23 23 4 12 12 3 I0 I0 1 6 8 0 6 6 0 l 11 50 50 50 46 40 30 23 12 i0 0 6 1 50 5O 50 46 4O 30 23 12 10 8 6 1 !