ISR11 Relevance Feedback in an Information Retrieval System chapter W. Riddle T. Horwitz R. Dietz Harvard University Gerard Salton Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. v'-53 Results for Query No. 9 QA9ANALYSIS Document Rank i[OCRerr]. Name Relevant Correlation 1 [OCRerr] 307R0LE OF / .00855386 2 714 305THE USE 0 .006014851 3 31 2O7PRACTICAL 0 14 148 8014CO[OCRerr]u[OCRerr][OCRerr]RI 0 5 57 603COMPILATI 0 6 15 6lOCHARACTER - .0171414216 7 82 3O8[OCRerr]CHI[OCRerr][OCRerr]E R X - .027261146 8 68 8llTHE RELAI - .02967852 9 13 l1l9RECEI'TT A - .03608157 10 18 ll2[OCRerr]TAIE-OF - .036314862 11 28 506THE ROLE - .03700040 12 114 5l8EDUCATION - .0140821614 13 So lOOlDESIGN 0 - .0141147700 114 25 209THE REDUC - .0141991400 15 7 21IA C0MPU[OCRerr]E - .014[OCRerr]7625 Ranking took .166657 seconds Relevant Document Ranks Document [OCRerr] Name Relevant Correlation_j 171821 308!4ACHI[OCRerr][OCRerr] R X - .027261146 [fiffil_1418ENGLISH-L x - .07926529 j Normalized Recall = .80214691 Normalized Precision = .1431451146 Rank Recall = .0857 Log Precision = .1313 Query(2) = Query(1) + Alpha N[OCRerr] ¶[OCRerr]I(lGHT) (I) * Document of Rank I 1=1 Document Rank 1 2 3 14 5 6 7 8 9 10 11 12 13 114 15 Weight 000000100000000 N = Number of retrieved documents Alpha = 2 (e) After First Update Figure 10 (continued)