Chinese Micro-Blog Sentiment Classification Based on Emotion Dictionary and Semantic Rules

Emotion dictionary is an important resource of micro-blog sentiment analysis. Based on the existing emotion dictionary, the paper improves the calculation method of emotion intensity of different emotion categories of emotion words in emotion dictionary. Firstly, we use word2vec to train the vectors of emotion words from the large-scale micro-blog text, select the benchmark words of each emotion category, and calculate the similarity between the emotion words vector and the benchmark words vector to obtain the emotion intensity of different emotion categories of emotion words. Then we analyze the influence of negative words and adverbs of degree on emotion tendency of different categories of emotion words, and define emotion semantic calculation rules. Finally, we combine with the emotion dictionary and the semantic rules to calculate the emotion intensity of different emotional categories of micro-blog, and realize the emotion classification. The experimental results show that our method is effective.