Fuzzy markup language with genetic learning mechanism for invention patent quality evaluation

Abstract

Patent quality evaluation and applications are important issues for new products generation. In recent years, Taiwan government also actively pushes the patent-related laws and rules to strengthen the development of national patent technologies and their protection utility. In order to deal with large amounts of patent information to enhance the patent expansibility and technology transfer possibility, this study proposes genetic fuzzy markup language (GFML) for patent quality evaluation. First, some patent gazettes are downloaded from Taiwan intellectual property office (TIPO) website. The GFML is used to describe the knowledge base and rule base of the patent's quality evaluation based on the evaluation index of Japan patent office (JPO) and intellectual property quotient (IPQ). Additionally, the patent quality evaluation ontology is also constructed. Then, we infer each patent's quality comprehensive evaluation based on the constructed ontology. We also adopt the genetic algorithm (GA) to improve the performance of the proposed method. Experimental results show that the proposed mechanism is feasible for the patent quality evaluation.