摘要
Abstract
[Purpose/Significance]User sentiment cognitive graph of online health community can reveal the cognitive mechanisms that contribute to user emotions and providing reference for in-depth analysis of user emotional needs,optimi-zation of community emotional support services,and development of health intervention strategies.[Method/Process]Firstly,the structure of user sentiment cognition graph of online health community was designed based on the OCC model;Then,the LSTM model was combined with a custom sentiment dictionary to extract user sentiment types and sentiment words,while the BERT-CRF model was used to extract sentiment triggers and their relationships such as event outcomes,subject behavior,and object features;Next,the extracted entities and their relationships were fused to construct a senti-ment cognitive graph.Finally,the user comment data of diabetes community was takan as a an example to construct user sentiment cognition graph and its functions were analyzed.[Result/Conclusion]The emotion cognition graph divides emo-tions into fine-grained categories from the perspective of cognitive science.It can not only be applied to analyze the emo-tional state of users,but also reveal sentiment triggers such as event outcomes,subject behavior,and characteristics of object things.The construction of sentiment cognitive graph can further improve the theory and methods of user sentiment analysis,while also filling the shortcomings of knowledge graphs in sentiment cognitive analysis.关键词
OCC模型/情感分析/情感诱因/认知图谱/在线健康社区Key words
OCC model/sentiment analysis/sentiment triggers/cognitive graph/online health community分类
医药卫生
