Research On Key Technology Of Patent Information Acquisition And Analysis System

Intellectual property with patent as a core has been gradually emphasized by increasing governments and enterprises. Nowadays, most of domestic enterprises adopt the relatively laggard methods in the aspect of patent resources acquisition and analysis, such as manual downloading and manual reading analysis, which is not satisfactory in efficiency and accuracy.This paper introduces the current situations of research on patent information acquisition and analysis at home and abroad. Related theoretical knowledge is also explained, such as concept retrieval, web page information extraction and social network analysis. The architecture of patent information acquisition and analysis system is proposed. According to the difficulties during the system construction, it does several researches on key technology and a protosystem is designed and implemented. The empirical research is about the retrieval of patents literatures related to semiconductor, which validates the operating effects of the protosystem.There are many improvements in accuracy, comprehensiveness and automation of patent information acquisition and analysis. The patent information sources chosen by protosystem are the authorized patent literatures in USPTO (The US Patent and Trademark Office).Major contributions of the paper are as follows:1. Concept retrieval is applied in the process of patent information acquisition. Domain knowledge base composed of WordNet semantic lexicon and DBpedia dataset can realize the semantic concept expansion for initial retrieval expression provided by users. Understanding and dealing with users'search request from the concept level will make retrieval process more intelligentized, which will also increase the precision and comprehensiveness of the search results.2. Social Network Analysis (SNA) method is applied in patent citation analysis. Pajek, a SNA tool, is integrated with the protosystem in order to make the visual display of the patent citation network possible. It intuitively displays the deep analysis results from patent literatures.3. The protosystem for patent information acquisition and analysis is built, including patent information retrieval subsystem, patent information extraction subsystem and patent information analysis subsystem. The basic functions of the protosystem are realized by virtue of concept retrieval, web page information extraction technology, SNA and so on. The paper selects the retrieval of patents literatures related to semiconductor as an empirical research, which validate the operating effects of the protosystem. Compared with traditional manual ways, the protosystem based on the Internet has superior performances in automation, efficiency and accuracy of patent information acquisition and analysis. The protosystem implements the whole process from patent information retrieval, acquisition to analysis with higher practical applicability, which provides crucial reference resources for patent strategy making.