IRS13 An Analysis of the Documentation Requests 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. X-33 1 5 Initial Search Relevance Feedback 6 10 First Iteration, Five New Documents i[OCRerr]amined 11 15 Second Iteration, Five New Documents Exandned Precision .60 - .55 .50 .45 .40 .35 .30 5 .25 .20 .15 Is .10 .05 - I I I I I I I I I I 0 .10 .20 .30 .40 .50 .60 .70 .80 .90 1.00 Recall ADI Abstracts Thesaurus, "pseudo-Cranfield" Cut-off, Macro Evaluation Over 35 Requests. Comparison of Performance Effectiveness of Relevance Feedback wit[OCRerr] Manual Modification by Important Term Weight Increases Fig. 20