Research On 'Topic+View' Extraction Method Based On WSO-LDA For Micro Blog Topic

With the popularity and development of computer networks,the Internet has entered web2.0,a new era of web services to user-centric,resulting in the rapid expansion of information,data significantly increased.In recent years,a large number of researchers are engaged in the research of text mining from huge Web data.Microblog is an information exchange platform with rich media information.As an important channel for the development of public opinion,microblog topics include a large amount of text content on the one hand and the user's emotional information on the other.Mining Webo topics of the topic of information and view content,and with a simple Weibo the model will reduce the dimension of the topic has important research significance.However,the existing text mining models and methods can not effectively mine and display microblog topics,so this paper presents a new text description model and corresponding mining algorithm.This paper first studies the text extraction theory,the emotion analysis technology and the theme model theory,and combs the related theories as the research foundation of this paper.On the basis of existing LDA model for mining the topic and view information of microblog topic at the same time(Weibo Sentiment Online-LDA,WSO-LDA)model was proposed and used to model the micro-blog topic.Based on the WSO-LDA model,this paper designs a topic-view extraction algorithm for micro-blog topics,and presents topic mining results from the two-dimensional text description model and the "theme + view" Finally,this paper uses Sina microblogging crawler tool set collected 5 different micro-blog topic Sina microblogging according to the published time to build experimental data set for comparative experiments,experiments show that the proposed model and algorithm is effective and reliable.The specific innovation points are as follows:(1)Sina Weibo theme characteristics,this paper mining micro-Bo topics in different time slices under the content and the evolution of view,this model is based on the onlinelda model added emotional factors,proposed the construction of TSLDA model.(2)A text description model of micro-blog topic is proposed,which describes topic information from two dimensions of topic and view.The topic-view extraction algorithm is designed to extract feature words from micro-blog topic.