Study On Expansion Of Sentiment Lexicon Based On Information Fusion

With the vigorous development of the e-commerce industry,how to quickly and accurately analyze the consumer's likes and dislikes of the product has become an increasingly important issue in business product decision-making.The tendency of product reviews based on emotional analysis is widely studied.But for clothing such a multi-category,the user is highly sensitive to the relationship between the product review information mining only concerned about the positive and negative polarity of the comment can not meet the emotional needs of clothing products,supplemented by emotional emotional evaluation of emotional polarity analysis can better support Decision of such products.In view of this demand,this paper studies and designs an emotional dictionary that combines emotional intensity and emotional polarity and applies it to emotional analysis of apparel products.The main work of this paper is as follows:Based on the analysis of emotional analysis and emotional dictionary construction technology,the structure of emotional intensity dictionary is studied.Aiming at the ills of the concentration distribution of the intensity of the emotional intensity dictionary,an intensity uniformization algorithm is proposed.On the basis of analyzing the classification of the existing emotional polarity,this paper studies the emotional polarity classification method which is suitable for constructing the emotional polarity intensity dictionary through information fusion,and constructs the emotional polarity dictionary.By translating the English corpus ISEAR,the Chinese corpus is formed,and a mapping mechanism based on this emotional intensity and emotional polarity is proposed to realize the emotional polarity intensity dictionary of fusion strength information and polarity information.Finally,the emotional polarity dictionary is used to analyze the emotion in the clothing product review data set,and the validity and value of the method are verified by comparison with the mature method.