An Analysis Of The Tendency Of Microblogging Text Of Fusion Emoticons

Nowadays, with the rapid development of the Internet and new media, microblog hasshown explosive growth, and more and more people have began to use it. On microblog, peoplerelease real-time news to express their views on various issues in the real life, discuss thecurrent hot topics, and share information and resources, etc. Due to mutual concerns,forwarding and comments between users, microblogs are rich in information, including largeamounts of emotional information. Sentiment orientation analysis on the information ofmicroblogs, accurately exploring people’s views on hot topics and products are of greatsignificance to early warning and network public opinion analysis, product research andmarketing management, etc.Existing Chinese orientation analysis mainly concentrate in areas such as product reviews,news reports, and so on. However, as a new social network media, the orientation analysis ofmicroblog still adopts the traditional orientation analysis method, lacking consideration ofmicroblog characteristics. In view of microblog characteristics, the work of this paper mainlyconsists of following points:1. Aiming at the language characteristics of microblog short texts, this paperproposes an improved N-Gram method for microblog new words discovery integratingmutual information.The language of microblog is so active and colloquial that there will a lot of new wordsevery day, which are often with a certain emotional orientation. Aiming at the characteristics ofthe new words in microblog, this paper proposes a method for microblog new words discoverywhich integrating mutual information and N-gram. It draws character strings from themicroblog corpus as candidate feature words, and uses mutual information to merge the featurewords that consist of several characters, so as to discover the new words in microblogs. Theexperimental results show that the proposed method can be applied to microblog new wordsdiscovery.2. According to the characteristics of the emoticons in microblogs, we build themicroblog emoticons dictionary.As a widely used network language, emoticons have become an essential way when peoplecommunicate in microblog. Aiming at the characteristics of microblog, this paper proposes a method foremoticons dictionary construction based on microblog statistics. We carry out text orientation analysis on thetext which co-occurs with emoticons, thus to determine the orientation of emoticons, to construct anemoticons emotional dictionary for microblog orientation analysis. The experimental results show that theuse of emoticons emotional dictionary improves the accuracy of microblog orientation analysis and achievesbetter classification effect. 3. On account of the characteristics of Chinese microblog, we propose an algorithmfor Chinese microblog orientation analysis which integrates emoticons.Based on the new words discovery algorithm and the built microblog emoticons dictionary,this paper proposes an algorithm for microblog text orientation analysis which integratesemoticons. The elements with emotional orientation and related grammatical features areregarded as emotional orientation information. In view of the users’ habits and microbloglanguage characteristics, we add emoticons and network new words to the traditional emotionaldictionary which mainly consists of emotional words, degree adverbs and negative words, inorder to effectively improve the accuracy of the microblog orientation analysis. In addition, thisalgorithm conducts the syntax analysis and syntax analysis to determine the logical relationshipbetween words and clauses in microblogs.