Intelligent recommendation methodology and system for patent search

Abstract

Patents are critical intangible assets for enabling an enterprise's creation and technology. Patent knowledge management is a time consuming task that often dominates the valuable time of R&D staff. This research develops an intelligent recommendation methodology and system for efficient and effective patent search. The system provides an algorithm which is an automatic search engine with management modules. The system clusters users' patent search behavior and infers new patent recommendations. The proposed methodology evaluates the filtered information of the searched patents. Afterward, the system clusters the users and finds each user's neighbors based on the collaborative filtering mechanism. By clustering neighbors' behaviors, the proposed system recommends new users to more appropriate patents to study. When enterprises are planning R&D policies or searching patents for possible prior arts, the intelligent recommendation system helps identify and recommend comprehensive and related patents while saving time and costs.