Using Summarization Techniques on Patent Database Through Computational Intelligence

Abstract

Patents are an important source of information for measuring the technological advancement of a specific knowledge domain. The volume of patents available in digital databases has grown rapidly and, in order to take advantage of existing patent knowledge, it is essential to organize information in an accessible and simple format. The classification systems groups, made available by patent offices, were given names capable of representing them and facilitating the process of searching for the information associated with its content. The purpose of this paper is to use automatic text summarization techniques to develop an automatic methodology to help the examiner to name new patent groups created by the categorization systems. We used three summarization strategies with two different approaches to choose the most representative sentence for each subgroup. The experiments were performed on the basis of abstracts and descriptions of patent documents, in order to evaluate the performance of the methodology proposed in different sections of the patent document. Validation experiments were conducted using four subgroups of the United States Patent and Trademark Office, which uses the Cooperative Patent Classification system.