IRS13 Correlation Measures chapter K. Reitsma J. Sagalyn Harvard University Gerard Salton Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. IV-13 as before. The above discussion is necessary when describing the modifications made to use the P-R-N formLlla for document - document correlations. With weighted, document descriptions vectors the formula becomes 7[OCRerr]ixi P-R-N = 7 (v[OCRerr]) 2 + 7 ([OCRerr]w[OCRerr])2 - 7vi[OCRerr][OCRerr]i where the summations are taken from i = 1 to d , and where d equals the number of concepts in the description vector. The interpretation of this function is not simple. The closest meaning which can be attached to the denominator is that it represents twice the maximum weight of the inner product of the two vectors, assuming perfect correlation, minus the actual inner product. The difference, there- fore, will always be greater than or equal to the actual inner product. By the argument previously used to determine the range of the binary P-H-N function, the range of this function can be shown to be from 0 to 1 , inclusive. G) The Stiles Coefficient The Stiles function incorporates the parameter b as it was defined for the [OCRerr]ron-Kuhns function. The formula is n( Inbi - n2 (7vi)( Twi) (n[OCRerr] 7+ (n 7wi) Since the formula was originally proposed to calculate index term - index St =