Research on Intelligent Patent Classification Scheme Based on Title Analysis

Abstract

With the rapid increasing number of patents, it is becoming more significant but difficult to mine underlying information from huge patent data in database. By integrating Latent Dirichlet Allocation (LDA) topic mode with text mining algorithms, we propose two patent classification schemes: topic-based patent classification and title word-frequency-based patent classification, which can be applied in the areas of patent retrieval, patent evaluation and patent recommendation. The process and implementation methods of proposed schemes are discussed, and the examples to intelligently classify patent records in the area of railway transportation in international patent database are given, the results can adequately verify effectiveness of our proposed schemes.