TOPIC DETECTION FOR EMERGENCY EVENTS BASED ON FCM DOCUMENT CLUSTERING
This paper discusses the usage of document clustering methods for topic detection of emergencies. Its main contribution is to apply the named entity of event-based framework to extract the feature terms of Web documents, exploit the TF-IDF method to weight the Web document characteristics of emergencies, and finally detect the hot topics through the FCM clustering algorithm. This method can reduce the redundancy feature terms of Web documents for emergencies effectively, and explore the internal structure and connections of the original data. It can also decrease the feature dimensions to improve the intelligibility of document data and the accuracy of topic detection to a large extent. Experimental results show that the FCM clustering method can achieve the topic cluster aggregation in the Web document sets, receive excepted topics of the Internet information sources timely, and monitor its related reports.
document clustering Topic Detection Emergency events Feature extraction
Tian Gao Junping Du Su Wang Liping Chen
Beijing Key Lab of Intelligent Telecommunication Software and Multimedia,Beijing University of Posts Beijing Key Lab of Intelligent Telecommunication Software and Multimedia, Beijing University of Post
国际会议
北京
英文
1181-1185
2010-10-26（万方平台首次上网日期，不代表论文的发表时间）