Analysis of Patents Related to COVID-19-Based on Patent Clustering Model in Specific Fields  

作　　者：Fu Nan Li Qian Yuan Hongmei 

机构地区：[1]School of Business Administration,Shenyang Pharmaceutical University,Shenyang 110016,China

出　　处：《Asian Journal of Social Pharmacy》2024年第4期371-382,共12页亚洲社会药学（英文）

摘　　要：Objective To improve the efficiency of patent clustering related to COVID-19 through the topic extraction algorithm and BERT model,and to help researchers understand the patent applications for novel corona virus.Methods The weights of topic vector and BERT model vector were adjusted by cross-entropy loss algorithm to obtain joint vector.Then,k-means++algorithm was used for patent clustering after dimension reduction.Results and Conclusion The model was applied to patents for corona virus drugs,and five clustering topics were generated.Through comparison,it is proved that the clustering results of this model are more centralized and the differentiation between clusters is significant.The five clusters generated are visually analyzed to reveal the development status of patents for corona virus drugs.

关 键 词：corona virus patent clustering patent analysis BERT model 

分 类 号：R95[医药卫生—药学]