Design And Implementation Of Patent Information Management System Based On.NET

The patent information is an important intelligence resource which contains awealth of technical knowledge. To fully exploit the patent information, to helptechnical staff to find new areas of technology and technical means is of greatsignificance to look for new economic growth point, to analyze the development ofthe industry and to formulate development strategies for companies. However,because of the vast number of patents and its fast growth, manual analysis of patentinformation is very difficult.There have been a lot of patent information acquisition and analysis softwares inwestern countries. But most of them only focus on English patents. In China, patentinformation is not well studied yet. In this paper, using the mainland patent in Chinaas the research object, we studied patent information acquisition and analysistechniques and developed a.NET-based patent information management system.The patent information in this paper is collected from the patent database of theState Intellectual Property Office website. According to the characteristics of this site,we designed and developed a patent information acquisition module. This modulesends a request to the site database. Firstly, obtain the source code of the searchresults page of the search criteria, then according to this source code using regularmatching method to parse out the address of each patent, thirdly, according to thisaddress send a request to the site database to obtain the source code of patent detailpage, fourthly using regular matching method to parse out the basic information andthe address to download instructions of the patent, finally, according to the address todownload instructions.In this paper, the patent information analysis is divided into two parts of the basicstatistical analysis and patent clustering. The research object of the basic statistical analysis is patents of the Ocean University of China. Basic statistical analysis isincluding the applicant analysis, analysis of the inventor and the agency. Throughstatistics patent the basic information, to reflect the applicant’s annual filings, thedistribution of annual grants and patent category, and the number of annual inventors,applications of the inventors, inventor patent weight, the number of the agency agent,regional distribution of the agency, etc. Under these circumstances, obtain the law ofthe applicant’s application and scientific research and technology strength. In thissystem patent clustering is including fast clustering and clustering and comprehensiveclustering, exampled by fast clustering in this paper. The research object of fastclustering is invention patents of the Ocean University of China applied in2011.Afterprocessing of name and summary information of these patents, use K-MEDOIDSclustering algorithm to cluster these patent texts. The user can set the value of K tocluster these patents. Besides, this paper provides a method to obtain the best value ofK according to the weighted similarity ratio of between-class and in-class to help theuser to select the most appropriate value of K.This paper presents a method of calculating the patent weight of inventor, is aninnovative of the analysis of patent inventor. This system takes full account of thefunction comprehensive, practical and quick, which is simple and friendly surface.This system is very wide application prospect and has a high practical value,especially for the analysis of a single applicant’s patents.