A Detection Method of Online Public Opinion Based on Element Co-occurrence
Discovering and identifying public opinion timely and efficiently from web text are of great significance.The present methods of public opinion supervision suffer from being rough and less targeted.To overcome these shortcomings,this paper provides a public opinion detection method of network public opinion based on element co-occurrence for specific domain.This method,considering the nature of public opinion,represents three main factors(subject,object and semantic orientation)that constitute public opinion by employing their feature words,which can be dynamically combined according to their syntagmatic and associative relations.Thus,this method can not only generate topics related to public opinion in specific fields,but also identify public opinion information of these fields efficiently.The method has found its practical usage in “Language Public Opinion Monitoring System and “Higher Education Public Opinion Monitoring System with accuracies 92%and 93%respectively.
Element co-occurrence Online public opinion Syntagmatic relations Associative relations
Nanchang Cheng Yu Zou Yonglin Teng Min Hou
National Broadcast Media Language Resources Monitoring and Research Center,Communication University of China,Beijing 100024,China
国际会议
广州
英文
322-334
2018-11-23（万方平台首次上网日期，不代表论文的发表时间）