Research And Application Of Patent Information Extraction And Topic Mining

Patent data is a very important source of scientific and technological information,it is human invention and creation,the expression of technological innovation,it represents the latest level of technological development,Therefore,it is of great significance to tap the hidden value of patent data.The traditional method of patent information analysis is mainly for the statistical analysis of the external characteristics of the patent,the amount of information obtained is limited,and the subject can describe the content information of the patent well;The effective words are an important description of the results of the patent,and it is of great value and practical significance to combine the patent theme,the effective words and the patent features into the research and application.In this paper,we use the technology of natural language processing,data mining and statistical analysis to study the topic of patent and the recognition of effective words,then,the paper analyzes the competition relationship of the organization with the combination of the subject and the patent agency,based on the combination of effective words and patent features,the paper puts forward the algorithm of patent sort,at the same time,the trend analysis of the effective words is also carried out,finally,the system of patent information extraction and topic mining is realized.The research contents of this paper include:(1)In the aspect of effective word extraction,a method of extracting the effective words based on the conditional random field model and the dependent syntax analysis model is realized,and the effective words in the patent data are effectively extracted.(2)In the aspect of subject mining and institutional competition analysis,a competitive analysis model of patent agency based on subject similarity is proposed,and the LDA theme model is extended to construct the themeinstitutional model,and the subject and the competition relation of the patent institution are carried out analysis.(3)In the aspect of patent sorting and subject function application,a sorting method based on patent efficacy intensity is proposed.The method solves the problem of subjectiveness of most indexes.At the same time,it studies patent retrieval based on efficacy words,the analysis of the theme efficacy trend and analysis of the relationship between the theme and the effective words and the inventor(4)In the engineering practice,a patent information extraction and theme mining system is designed and implemented.The system is divided into data maintenance and retrieval module,multi-dimensional analysis module of effective words and subject and institutional competition analysis module.Finally,the relevant experiments and comparative analysis are carried out on the patent data set in the field of new energy vehicles,which proves the feasibility and effectiveness of the scheme and provides a solution for the effective mining of patent information.