Research On Building Visual Analysis System Of Patent Information

Patent information has covered more than90%scientific research achievements of theworld and80%of the technological achievements is not included in the academic and publicpublications. According to the World Intellectual Property Organization, the global patent hasreached63million currently. Generally, the ordinary people are hard to find the distributionand implicit information of patent from so many patents with naked eye, which needs with theaid of modern information technology and professional patent analysis tools. Because patentanalysis system at home and abroad have different limitations and use the high cost, thisarticle introduced low-cost solutions of patent information visualization and implement it.Firstly, this article puts forward the concept of patent information visualization analysissystem from the shortage of the existing patent analysis system. Then the paper introducesand demonstrates construction and implementation process of the patent visualization analysissystem from several aspects of the data collection, data warehouse building, ETL designing,multidimensional analysis model and data mining model building. The article introduces andexpounds in detail building process of the patent data warehouse which is the physical storageof data and designs the four fact-tables and four dimensions. Although later sections discussonly two of the fact table, the remaining is follow-up research and development of resources.ETL system cleans and extracts the patent data collected and loads this data into the patentdata warehouse. The design of data warehouse and ETL system design will take50%~60%’stime of the whole project. Multidimensional analysis model, subsequently, designing processis described from the angles of the dimension and data cube building. After builtmultidimensional analysis model, several indices of patent bibliometrics and patent analysiswhich base on the model are added. To extend the system more, the correlation analysis, andcluster analysis and time series analysis model for data mining is introduced. Finally, thesystem has been carried on the simple application after completion of it.This article has two main innovation points. For one thing, BI is used in the patentinformation visualization analysis. Secondly, a low-cost and easy to use patent informationvisualization analysis solution is set up and using the patent data of Hunan province carriedon the empirical analysis and effective inspection.