Patent valuation based on text mining and survival analysis

Assessing the value of a patent is crucial not only at the licensing stage but also during the resolution of a patent infringement lawsuit. In this study, we use text mining to identify important factors associated with patent value as represented by its survival period. The variables retrieved from text mining were the Euclidian distance of patent claims between a patent and its backward cited patents, or forward cited patents, and the singular value decompositions (SVDs) of patent claims. After applying Weibull regression to 3D printing patents, the following factors were found to have significant associations with the survival time of a patent: the Euclidian distance of claims between a patent and its forward cited patents, the average number of forward citations, whether Germany is included among the family patent countries, whether the patent is transferred to another, and the presence of five SVDs. Our study is expected to contribute to enhancing patent valuation in consideration of patent infringement risk.