Game Theory Based Patent Infringement Detection Method

With fiercely increasing competition of intellectual property, the protection of intellectual property (IP) is paid increasing attention from worldwide. Effective patent infringement detection is the foundation of patent protection. Given the big patent data, manual patent infringement detection is inefficient and error prone. The design of automatic infringement detection method faces some challenges including: (1) how to detect patent infringement by novelty and non-obviousness; (2) which parts of a patent are selected against those of its counterpart patents in infringement detection. To solve the above issues, a game theory based patent infringement detection method is proposed. In the method, both the novelty and the non-obviousness are considered when the patentees take actions. The infringement detection is a game process to find the best action. Our method is compared with the state-of-the-art method on some patent data sets. The results show that the dynamic game method outperforms the baseline method in the evaluation measurement. Such method can be applied to patent examination and patent infringement litigation.