Mining And Analysis System Of Network Porpularity Imformation

With the development and popularization of the Internet,the network has replaced the traditional media from which people get their concerned content,news and other imformation.Attendantly,imformation popularity analysis in Internet become more and more important.The occurrence and spread of network information will bring a series of social impact.It's source is wide,it's propagation time is short,it's spread speed is fast,all these reasons increased the complexity of the relevant studies.Based on this demand,a set of functions and analysis system is designed and implemented in this article.Firstly,in order to simplify the subsequent elaboration of the sysytem,This paper briefly describes some of the key techniques and algorithms model which is involved by the system.Additionally,the demands of the system is splited to functional demands and non-functional demands.Taking this as a starting point,designed the overall stucture,explanations and demonstrations are given to illustrate each sub-structure.Mean whi le,implemented all the modules In strict accordance with the design.When implementing,this paper compared the candidate options and explained the reasons.Finally,designed the testing indicators and proved the excellent effect of mining and analysis system of network porpularity imfomation.In the system's design and implementation process,this paper presents the theme of dimensionality reduction and clustering algorithm based on LSA and PLSA to solve the problem that the data doesn't match the dimension,optimized the inappropriate part in the original algorithm.Unlike the topick models which are presented above,dimensionality reduction clustering algorithm ensured low loss at semantic imformation when decreased the amount of the redundant oreprations,improve the stability and performance of the system significantly.In addition,the paper processed a set of related mathematical derivation to make the porpularity analysis result displayed more intuitively.