Research And Implementation Of Microblog-based Sentiment Analysis System

In recent years,with the rapid development of the Internet,social networking site has become the main platform for people to express their views.Micro-blog,one of the most popular platforms in social network,generates a huge amount of user behavior data of great research value in many fields.Sentiment tendency analysis is one of the hottest research fields which analyzes user behavior data using statistics and machine learning methods,and then predicts user's sentimental attitude according to the analysis result.This thesis mainly studies and implements the sentiment analysis system for micro-blog text and it includes the following six aspects.First,several sentiment tendency analysis algorithms are studied,including support vector machine,naive bayes,Adaboost and neural network.Then it studies the principle of the three algorithms and compares the difference between them.Second,the web layout of micro-blog platform is investigated and a distributed micro-blog crawler system which crawls hot topic data from micro-blog platform,including micro-blog contents and comments,is designed.Third,a data preprocessing system is constructed,which uses some rules to process data.Fourth,the characteristics of micro-blog text are analyzed,and then the text feature selection methods are chosen.It uses chi-square and TF-IDF methods to select text features.Fifth,a text classifier is designed using the first three classification algorithms mentioned above and micro-blog texts are divided into three categories,including positive,negative and neutral.Meanwhile,the results of these algorithms are comparatively analyzed.Sixth,it also designs and implements a display system that fetches topic data from database and shows data on web page.Finally,based on the micro blog topic data,this thesis tests the sentiment analysis system.The results show that the system has a good effect in the micro blog sentiment prediction.