A method for assessing patent similarity using direct and indirect citation links

Assessing patent similarity is a fundamental and critical step in patent citation analysis. When evaluating a similarity among two patents, considering both direct and indirect citation links leads to more precise similarity assessment. This study proposes a method for assessing patent compound similarity that includes direct and indirect similarities. Given a direct similarity matrix that represents a patent citation network, the method calculates indirect similarity matrices and then obtains a compound similarity matrix. Keyword analysis in the text mining is employed to obtain a similarity for a pair of patents. In addition, two criterion are proposed for validating the compound similarities for the patent citation network.