Research on Model and Algorithm of User Access Pattern Data Mining

The era of big data has come. Today, all kinds information and are showing explosive growth. Internet activities different scales struggling to catch up with the pace development "big data. This paper uses time series mining algorithms, Microsoft's tools model sets collected from halls, so as discover user's online behaviour patterns potential rules within a certain period time. we made reasonable suggestions for scientific management campus network. A new clustering method based on improved Kohonen self-organizing feature mapping neural network is proposed. Gaussian-shaped membership function introduced output several neurons degree greater than threshold, thereby solving problem users' multiple interests.