A Study On The Impact Of Patent Network Characteristics On Green Patent Value In The Case Of LED Industry

In recent years,the growing problems of environmental pollution and the shortage of resources have brought great influence to environment,which caused serious threat to normal productions and living conditions.In a variety of ways to deal with environmental changes in the exploration,the rise of green technologies and green patents provide new solution ways and directions.However,the green patent applications require invest a lot,meaning that if the quality of the technology is inferior,too many patent applications will lead to an increase in business costs,which cause a negative effect to enterprises.So,when faced the increasing patent applications,it is essential for enterprise to master the method of assessing the value of patents.There are many literatures to study the value of patents,but mostly from the perspective of patents’own attributes,such as patent forward citations,backward citations,patent claims and so on.Nevertheless,patent value is not static,it is often different in the process of technical knowledge flow.Social network analysis can present patent information from a dynamic perspective by using the relationship between the patents to build a technical network.In order to explore the patent value,this study introduces the patent citation network analysis and combines the patents’ own attribute indexes.LED is selected as the representative industry in the field of green technology and the LED patent data from Thomson Innovation database is selected in this study.It includes the US authorized patent information up to May 31st,2011.According to the number of patents H index,the patents are divided into high cited patents and low cited patents,and in accordance with the principle of 1:2 data screening,the final 264 patents were included in this analysis.Among them,the high cited patent area includes 88 LED patents,and the low cited patent area includes 176 LED patents.Then,the study uses the different patent category as the proxy variables of patent value,and the social network indexes,out-degree centrality,out-closeness centrality,eigenvector centrality and effective size as independent variables,the patent claim counts and inventor counts as control variables to construct the Logistic regression model.The results show that out-degree centrality and effect size have significant influences on the patent value.