Research On The Evaluation Of Patent Value Under The Network Platform Service Economy Environment

With the rapid rise of patent-centered intellectual property trading network platform with patent as the core,the key to improve the activity of platform patent trading is to solve the accurate evaluation of patent value under the economic environment of network platform service.At present,most patent trading platforms still use traditional economic methods and comprehensive evaluation methods to evaluate patent value,but the practice and effect of evaluation are not ideal in the face of a large number of patent data.In addition,the research on patent value evaluation mostly emphasizes the influence of patent's own factors on the value,ignoring the external environment of network platform services.And the existing evaluation index systems lack certain theoretical guidance,are relatively scattered.Intelligent patent value assessment methods are still in the preliminary stage of research,and the existing models lack the screening and utilization of effective patent information.In view of this,based on the value chain theory,this paper established a four-dimensional evaluation index system of patent value under the network platform,which includes technical value,legal value,use value and platform service value.It comprehensively considers the impact of basic patent activities and auxiliary activities of network platform services on the value-added of patent.In addition,this paper proposed an intelligent patent value evaluation model based on grey correlation analysis(GCA)and random forest regression(RFR).The model uses the grey correlation analysis method of entropy modified G1 to screen high correlation patent samples,and then trains the random forest regression model to realize the accurate evaluation of patent value by revealing the non-linear relationship between patent transaction price and each evaluation index under the network platform.Empirical research shows that the GCA-RFR model is feasible and applicable in the practical application of patent value evaluation of network platform.Compared with the traditional BP neural network model,it has higher robustness and can show better forecasting effect in the case of limited sample size,and has higher accuracy compared with the conventional RFR model.The paper enriches the theoretical research of patent value evaluation method,and puts forward an intelligent method of patent value evaluation suitable for the economic environment of network platform service,which can provide scientific reference for platform patent pricing,so as to improve the patent transaction rateand promote the development of patent transaction platform.