A Method To Track Topics Based On Events Of Micro-Blog

In an increasingly advanced information age,micro-blog as a product of the information age,it has become an important public carrier and route of transmission.Recently as the main stream of social platform,micro-blog has many functions,and information dissemination can be the most crucial role.The content information of Micro-blog could spread among different users,and when the original author of a micro-blog publish his own content information in the social platform,other micro-blog users not only could forward it directly,but may modify and then publish it.Therefore,given the characteristics of micro-blog information dissemination,when the content information of micro-blog is published,it may be forwarded and modified among different micro-blog users,which leads to a problem,how to judge the change of content information and the impact of the public because of content information changes,and then needs to be solved by finding the dissemination source of micro-blog information.Thus mastering the route of information dissemination and then tracing the dissemination source of micro-blog information,the media of network could be controlled effectively and the effect of micro-blog could be maximized.Firstly,according to the characteristics of micro-blog text and that similar values are not ideal for the space vector model revealed text,and poorly semantic association;the accuracy of LDA model is influenced by unsupervised modeling method,and data noise,this dissertation proposes a new model for this research,namely,the space vector and LDA-based interaction model.Secondly,based on the characteristics of micro-blog text and that basic ant colony clustering algorithm does not set any text scale two-dimensional grids and deals with isolated and discreted data objects casually,this dissertation proposes a improved ant colony clustering algorithm.Then,based on the above text interaction model and the improved ant colony clustering algorithm,this dissertation tracks topics of micro-blog event,and detailedly describe the whole implementation process of the method of tracking topics based on micro-blog events.Finally,by simulation experiments,this dissertation verifies the method of tracking topics based on micro-blog events;the simulation experiment is divided into three parts,namely,the experimental analysis of text modeling,clustering algorithms and tracking topics based on micro-blog events;in each experimental section,this dissertation presents the experimental results by graph,chart and comparative experimental analysis,and the results of these experiments are expected which are in line with the target of this dissertation at the same time.