Research On The Construction Of Emotion Thesaurus And Algorithms Of New Cyber Words Identification

With thedevelopment of e-commerce and social network, microblog, one of themost popular media, is not only an important platform for individuals to exchange ideas,chat and deliver comments, but also an advanced position for enterprises to identify themarket needs, occupy the market share and compete with other companies. Thereforeuseful subjective information generated by microblog is becoming more and moreimportant to meet the customer demand, predict customer behavior, analyze useropinions and process human-computer interaction, etc. In order to solve this practicalproblem, a key technology of public opinion orientation analysis is proposed. Byhandling the text classification, text clustering, text retrieval or information extraction,we are able to extract the useful information among the big data of the internet. Andthen large of emotional information which reflectsthe mainstream viewpoints,commercial development and social behavior trend can be figured out to satisfy allkinds of people.The new research suggests that the Chinese emotion thesaurus is the basis resourcefor analyzing the emotion tendency of Chinese words. And the quality of the emotionthesaurus is closely related with the outcome of emotion orientation analysis. However,the current emotion thesauruses are not satisfactory: either with too few words or withinaccurate classification, or with monotonous polarity description. Besides, with thewide application of the internet, there are many net-words emerging rapidly and someold words being given new meanings, which raises the new demand of emotionthesaurus. Hence, how to find out the new words or old words with new meanings is acrucial challenge to construct the emotion thesaurus. Focusing on the problems above,the main research contents and creativities of this paper are showing below:This paper introduces the design and implementation of the emotion thesaurusbased on the ontology. Compared with the current Chinese emotion thesaurus, this paperimproves the system structure of the emotion thesaurus and proposes the encoding rules.And it also highlights the fine-grained emotion classifications which makes it muchcloser to the emotional tendency of human beings. Meanwhile, the thesaurus is able toupdate with an automatic algorithm. Finally, the feasibility of this method is proved bythe experiment. This paper proposes the algorithm for discovering the old net-words which havenew meanings. Based on the emotion thesaurus we have constructed, this paper focuseson the new net-words in microblogs which would lead to semantic fuzziness. And itproposes a three-step net-words algorithm to identify the words with new meanings.Finally the result of the experiment has proved the validity of the algorithm.