Research On Micro-blog Hot Topic Detection And Tracking

With the coming of information era, the internet has become an important channel forpeople to access, publish and disseminate information quickly. Micro-blog followingbroadcast, newspaper, TV, becomes a significant role of information dissemination. Fromthe first Chinese social networking FanFou lunched in2007to China micro-blog winningsupport among the people in2009, micro-blog has become the mainstream internetapplication and is increasing rapidly. The big data and rapid propagation make grasping theuser’s concerning more and more difficult. As a consequence of the phenomenon, the taskof hot topic detection absorbs the attention on the stage. The user can only receive theinformation published by their followed users, but this results in the local informationphenomenon and the can not understand the full view of what they caring. Althoughmicro-blog platform provides a hot topic list application, the manual editing leads to topicnot objective enough. Therefore, this paper studies micro-blog detection and tracking havecertain practical value.First, we introduce the concept of topic detection and tracking technology, and thenwe have the related clustering and classification algorithm. In this paper, we propose anovel topic detection method which combines with the characteristic of micro-blog,semantic topic mining method and get a better result in accuracy. Meanwhile, we can makea hot topic list supported by our topic detection results; In addition, this paper presents anapproach to incremental classification for micro-blog topic tracking, and achieved goodresults. Finally, Experiments show that the algorithm is effective and accuracy.