文摘
A hashtag is an important metadata in microblogs and used to mark topics or index messages. With topic-related hashatags microblogs are well grouped, and users can retrieve the microblogs efficiently and then follow the interested conversations. At the same time, microblogging service providers can leverage hashtags to classify the massive microblogs for building high-level applications such as event detection and tracking, sentiment analysis, and opinion mining. However, statistics show that hashtags are absent from most of the microblogs. In this paper, we summarize the similarities between microblogs and short-message-style news, and then propose an algorithm named 5WTAG for detecting microblog topics based on the model of five Ws(When, Where, Who, What, hoW). Since five-W(5W) attributes are the core components in event description, it is guaranteed theoretically that 5WTAG can extract the semantical topic from a microblogs properly. We introduce the detailed procedure of the algorithm 5WTAG in this paper including microblog segmentation and candidate hashtag construction. We propose a novel method of recommendation computing for ranking candidate hashtags, which combines syntax analysis and semantic analysis, and observes the distribution law of human-annotated topic tags. We conduct comprehensive experiments to verify the semantical correctness and completeness of the candidate hashtags as well as the accuracy of recommendation using the real data from Sina Weibo.