A method for determining relevance of papers and patents

Abstract

The application discloses a paper and patent relevance determination method, and relates to the technical field of data processing. The method comprises the following steps: firstly, papers and patents with similar keywords are gathered into topic clusters to complete the structure conversion from the literature level to the topic level. Then, the keywords of each paper, patent and corresponding topic cluster are comprehensively determined to contain the papers of each paper topic cluster and the patents of each patent topic cluster, so that each paper and patent is explicitly divided into the corresponding topic cluster. Then, the paper-patent citation network, the paper-patent semantic network and the paper-patent keyword semantic network are utilized to determine the citation relevance, semantic relevance and keyword semantic relevance between the paper topic clusters and the patent topic clusters from the three dimensions of citation, semantics and keyword semantics. Finally, the relevance of the three dimensions is automatically considered to determine the relevance between the papers and the patents, and the identification efficiency and accuracy of the relevance between the papers and the patents are improved.

Images