Revision as of 16:16, 14 January 2014 edit Kzhilton (talk \| contribs) 3 edits →Definitions Tag: gettingstarted edit ← Previous edit		Revision as of 06:37, 5 February 2014 edit undo Jm pedoc (talk \| contribs) 1 edit mNo edit summary Tag: gettingstarted edit Next edit →
Line 4: ==Definitions== GVSM introduces a term to term correlations, which deprecate the pairwise orthogonality assumption. More specifically, ~~they~~the ~~are~~factor considered a new space, where each term vector ''t<sub>i</sub>'' was expressed as a linear combination of ''2<sup>n</sup>'' vectors ''m<sub>r</sub>'' where ''r = 1...2<sup>n</sup>''. For a document ''d<sub>k</sub>'' and a query ''q'' the similarity function now becomes: Line 12: where ''t<sub>i</sub>'' and ''t<sub>j</sub>'' are now vectors of a ''2<sup>n</sup>'' dimensional space. Term correlation <math>t_i \cdot t_j</math> can be implemented in several ways. AsFor an example, Wong et al. ~~use as input to their algorithm~~uses the term ~~occurrence~~occurence frequency matrix obtained from automatic indexing as input to their algorithm. The term occurrence and the output is the term correlation between any pair of index terms. ==Semantic information on GVSM==

Generalized vector space model: Difference between revisions