[Purpose/Significance]The effective integration of the multiple relations in the patent citation network is conducive to eliminating the influence of subjective motivation on the authenticity of technical relation and improving the objectivity and accuracy of technology evolution path identification results.[Method/Process]Taking the link connectivities,subject similarities and citing intentions in the patent citation network as the basic re-lations,strengthened relations and regulation relations,it proposed a main path analysis method based on multi-re-lation data fusion.On this basis,it constructed a technical evolution path searching model based on multi-relation data fusion,and solved the model by using an iterated local search algorithm.Then,it carried out an empirical re-search t in the field of graphene sensing technology as an example.[Result/Conclusion]The experimental results show that the proposed method can effectively improve the effect of technology evolution path identification in the empirical field.In the future,much work in main path of multi-relation fusion needs to be done in discovering and applying more types of theme associations with integration value,and strengthening the research and verification of data fusion algorithms and models.
关键词
主路径/技术演化/多元关系融合/专利引用网络/引用动机/局部搜索算法
Key words
main path analysis(MPA)/technology evolution/data fusion of multi-relations/patent citation network/citing intention/local search
维普
万方数据