A Predictive Model for Patent Registration Time Using Survival Analysis

The infringements and suits of patents have been increased in many technological fields. Since a patent is an intellectual property to protect inventors' exclusive rights of developed technolog ies, technological pre-occupancy is very important to the devel- opment of companies. For the technological pre-occupancy a patent registration time is surely important. In this paper we propose a predictive model of the patent registration time using survival analysis w ith a Weibull survival regression model and a Cox's propor- tional hazard model. They are modeled on the number of international patent classifications (IPC) codes and keywords. To verify the proposed models, we perform a case study using the retrieved patent documents of 'hybrid vehicle' from the website of United State Patent and Trademark Office (USPTO).