What factors go into computing Relevance scores?
Ocean Tomo PatentRatings relevance scores are quantitative representations of the technological relatedness of any one U.S. patent to any other. A relevance score fundamentally represents the probability (on a scale from 0 to 1) that the first patent is citationally related to the second (i.e. that one directly cites the other), independent of the actuality of their citational relationship. Relevance scores are calculated by analyzing up to six generations of citational information (i.e. one generation includes all patents cited by or citing the reference patent). The analysis essentially determines the degree of overlap between the citational relationships of the two reference patents, and treats more overlapping citational relationships as indicating a higher probability of the two being directly citationally related. The Relevance Engine™ module contains over 4.8 billion (as of December 2009) patent-to-patent relevance relationships above a threshold value of 0.05. Typically over 1,
