An Evaluation of Bibliographic Couplings in Clustering Patents

Clustering bibliographic couplings is one form of document clustering. The purpose of this research was to measure the precision and recall of this type of clustering for patents. Design/Methodology/Approach: 717 US patents cited in patents related to the class 977/774 of the US Patent Classification were clustered using bibliographic couplings and the fuzzy c-means clustering algorithm. Clustering patents using bibliographic couplings performs well and produces a high recall rate. Considering the desirable performance of clustering based on bibliographic couplings، using bibliographic couplings may also perform well in the process of organizing and retrieving patents.