Research On The Model And Platform Of Hotspot Detection Based On Micro-blog

With the innovative production and dissemination mechanisms, micro-blog has played an important role of the birth and transmission of hotspot. The real-time hotspots hidden in the"light-speeded micro-blog fragments"have created the micro-blogging platform a unique advantage. How to find hotspots from the vast amounts of "light pieces" has become an urgent demand of internet information.It was introduced in this thesis that mining real-time hotspot from the mass of the"light pieces"is of great importance on decreasing"lost in information"through literature review. The idea based on the"centralization framework"was proposed, in which micro-blog and its characteristics, network hotspot and its characteristics, hotspot detection mechanism, the application of centralized processing methods were included. Production and processing model of hotspot detection based on micro-blog was designed and analyzed. With the establishment of micro-blog's similarity measure method and centralized processing methods based on communication path and user behavior, the production and processing hotspot detection process from raw materials to fittings then to hot clusters was realized. According to the model, the hotspot detection platform based on micro-blog was designed, including the corpus centers, production and processing schemas, and production and processing process based on schema matching. With the simple and rich metadata, not only the threshold of technology was lowered, but also the hot cluster analysis based solely on deficiencies identified was improved, and also hotspot detection based on the specified classification condition was carried out. Finally, a real experiment based on the model and platform was built and the experimental results were analyzed to verify the theoretical results.