The Study Of Topical Relevance Analysis In Chinese Microblogging Environment

With the development of Web2.0, microblogs and other social media have become the main platform to help people browse the network information and keep in touch with the trend in society. Microblogs are published by individual users and peoper share the information through social interaction. Then those result in a large scale microblogging information, and varying quality of the microblogs, real-time updates of trending topics. So it is of important significances to extract the topical subjects and meaningful pieces from huge amounts of microblog data, and it also has great importance for Internet information processing to find out the useful information.In this paper,we firstly study the the topical relevance analysis between Hashtags. Hashtag is a kind of topic label in microblogging Environment. The publisher can use the tag words(Hashtag) label the topics of his microblog. And the microblogging platform clusters the massive microblog data base on the Hashtag to help users find hot topics. However, different Hashtags created by different publishers may describe the same topic. Thus mining the relevance between the Hashtags will help to find hot topics more efficiently. In this paper, a wide range of features were explored to analyze the topical relevance between Hashtags.Such as the Hashtag text, content of the related microblog, the time of occurrence and the co-occurrences of Hashtags. The experimental results show that the proposed features are helpful for topical relevance analysis of Hashtags.In order to extract meaningful information from magnanimity microblog data to help users quickly understand all aspects of a topic, we carry out topic and sub-topic keywords extraction. The microblog are clustered to find the sub topics and keywords are extracted by running a keywords extraction algorithm. Experimental studies show that topic and sub-topic keywords help users quickly get the general ideas of a microblog data set. And it also improves the situation that user could not obtain the full information of hot topics because of high rate of microblog update.