| http://www.w3.org/ns/prov#value | - As such, each Web page is represented as a feature vector, whose element is a word with its weighting xi=<xi1, xi2 . . . xin> The weighting xij is calculated by length-normalized log(tf).idf, which has the form: idf t ??? 1 + log ( f d , t ) 1 + log ( avef d ) ??? 1 avedlb + S ??? ( dlb d - avedlb ) , where, d represents the original document, t represents term, fx,t represents frequency of term t
|
