A New Method for Micro-blog Platform Users Classification Based on Infinitesimal-time
Nan Liu · Journal of Information and Computational Science · 2013
In this paper, we propose the concept of infinitesimal-time slice and build infinitesimal-time model for micro-blog platform user classification problems. After researching on users in each infinitesimal-time slices, we integrate the results from all time slices to get the universal user groups in the end. Meanwhile, in order to avoid the traditional method which only relies on micro-blog platform tags to build users’ behaviors network to study users classification, we import analysis of micro-blog content with natural language processing technology to extract user interest and form the user interest vectors as the main classification basis. And we also propose improved Bayes classification algorithm to experiment on the universal user interest vectors and get the final user groups. According to experimental results, we find that the micro-blogging users based on time slice classification method can effectively avoid problems of the large computational complexity for large data, and it will has a certain role to promote the micro-blogging users classification study.