The Key Technologies’ Research And Implementation About Information Acquisition And Emotion Classification On New Social Network Media

Text sentiment classification is to do the emotional tendencies classification of texts bycomputers.Text sentiment classification can be used to predict the network public opinion,to analyse the advantages and disadvantages of commercial products, and even can alsojoin the the content of sentiment classification in the on-line user behavior mining. Inrecent years, with the development of network technology and mature, the networktransition from computer-centric to a network user-centric, users can express their viewsand comments freely on the Internet, The appear of the microblogging has also speeded upthe communication between users on the network. A large number of microbloggingcomments flood the network. How to serve the public using these data effectively is moreand more urgent.In this paper, design a sentiment classification algorithm based on the vector spacemodel, it uses the emotional evaluation object and emotional evaluation unit of emotionalsentences to establish the emotional vector, calculates the emotional tendency ofemotional sentences, and draws the appropriate emotional tendency weight value. Andfinally a sentiment classification retrieval system of microblogging is designed andachieved. The works done in this paper are as follows.1) It analyses the data interaction process of Sina microblogging using the HttpWatchwhich is a data packet capture tool, and achieves the crawling of the Sina microbloggingdata.2) It filters the emotional dictionary and proposes a extraction method of emotionalevaluation unit in emotional sentences combined with emotional dictionary.3) It studies on the emotion weight calculation of emotion words using the regularityof normal distribution, designs experiments and analyzes the results.4) On the basis of the vector space model, it proposes a short text sentimentclassification algorithm, use the emotional evaluation object as dimension, use theemotion weight of the emotional evaluation unit as dimension weight,then with theestablished vector space model, calculate to gain the emotion category and emotion weightof emotional sentences.Finally, the experiments show that the proposed extraction method and sentiment classification algorithm are scientific and feasible.