SP500215 Effective and Efficient Retrieval from Large and Dynamic Document Collections chapter D. Knaus P. Schauble National Institute of Standards and Technology D. K. Harman 1. Exact and Approximate Retrieval Status Values "e[OCRerr][OCRerr]": RSV(q, d[OCRerr]) 1 I d[OCRerr]I [OCRerr] * Ua[OCRerr][OCRerr]o[OCRerr][OCRerr]: RSV0(q) d[OCRerr]) 1 a[OCRerr]95 * jd[OCRerr] 2. Document Length Id; AI[OCRerr]%2 3. Weights of Query Features : ff([OCRerr]j,q)*nsdf([OCRerr]i) altfl[OCRerr] : b1 [OCRerr] 4. Exact and Approximate Weights of Document Features "nic" : "inc" 0 a[OCRerr]5 ff([OCRerr], d[OCRerr]) max{!f([OCRerr]h, d[OCRerr]) `9h E d5} * nidf([OCRerr]) 1*TLidf([OCRerr][OCRerr]) (1 +log(ff(tp[OCRerr],d5))) (1 + log(max{ff([OCRerr]h, d[OCRerr]) I Edj})) 1 "itc" : := (1 + iog(f f([OCRerr],d,))) (1 + iog(max{ff([OCRerr][OCRerr], d[OCRerr]) I [OCRerr]h E d[OCRerr]})) a19[OCRerr] 1* nidf([OCRerr]) 5. Normalized Inverse Document Frequency * nidf([OCRerr]) log(1 + df(tp1)) nsdf([OCRerr]) 1 log(1+n) Table 1: Classification of weighting schemes. 170