SP500207 Combining Evidence from Multiple Searches chapter E. Fox M. Koushik J. Shaw R. Modlin D. Rao National Institute of Standards and Technology Donna K. Harman Figure 2: Phase 1, Disc 1, WSJ, queries 50-100, Ad Hoc Merge Top ranked evaluation vpidt2 all 50 documents over all queries 10000 4056 1371 ion Averages: 0.5153 0.2841 0.1898 0.1285 0.0988 0.0746 0.0527 0.0338 0.0271 0.0222 0.0222 for all points 0.1317 for 3 intermediate points (0.20, 0.50, 0.80) 0.0972 Run number: Num~queries: Total number of Retrieved: Relevant: Rel[OCRerr]ret: Recall - Precis at 0.00 at 0.10 at 0.20 at 0.30 at 0.40 at 0.50 at 0.60 at 0.70 at 0.80 at 0.90 at 1.00 Average precision 11-pt Avg: Average precision 3-pt Avg: Recall. at 5 docs: 0.0476 at 15 docs: 0.0640 at 30 docs: 0.0843 at 100 docs: 0.2050 at 200 docs: 0.4481 Precision: At 5 docs: 0.3160 At 15 docs: 0.2147 At 30 docs: 0.1760 At 100 docs: 0.1246 At 200 docs: 0.1371 6 Results of Phase 2 6.1 Base Runs In Phase 2, twenty-live base runs were made on Disc 1: 5 different retrieval methods were used for each of the 5 sub-collections. Based on our evaluation, the 11-point averages are given in Table 4. Note that in the p-norm case, document weights were utilized, in contrast to the binary weighting used in Phase 1. 324