A Clustering Algorithm for Heterogeneous Network and Its Application on Patent Analysis
For quite a long time, patent analysis focuses on single type of patent information, such as co-word relationship or co-citation relationship.However, such kind of analysis can only reflect a partial understanding but not the whole picture of patent information.For this concern, researchers began to analyze patent network consisting of multiple types of nodes and relationship, namely heterogeneous network, to get deep insight of patent information.In this paper we proposed a new algorithm for heterogeneous network clustering, namely PathselClus+.This algorithm modifies previous PathSelClus algorithm in three aspects to avoid its main drawbacks.Firstly, incapability of taking heterogeneous network with more than one independent components as input;secondly, irrationality in its assumption;thirdly, impossibility of optimizing the weight values of meta-paths.Finally we apply PathSelClus+ on heterogeneous network consisting of co-citation relationship and citation coupling one, the clustering result not only reasonable but also shows the degree of each relationships contribution to the final clustering result.
patent analysis heterogeneous networks EM algorithm
CHEN Liang LEI Xiaoping SHANG Weijiao YANG Guancan ZHANG Jing
National Engineering Research Center of Science and Technology Information,Institute of Scientific a Research Institute of Forestry Policy and Information, Chinese Academy of Forestry, Beijing 100091,
国际会议
武汉
英文
234-238
2016-10-21（万方平台首次上网日期，不代表论文的发表时间）