Emotional Visualization Analysis Based on Online Book User Comments

Targeting different online reviews from online book users, it is possible to achieve intuitive emotional visualization analysis and research. Obtaining user emotional tendencies is a key guarantee for improving the overall market competitiveness and online services of e-commerce websites. This article is based on online book reviews to mine and analyze user emotions, using the public dataset Book_ Review, using the NLTK tool for data cleaning and natural language processing, using the TF-IDF (Term Frequency Inverse Document Frequency) algorithm model to achieve sentiment classification of comment texts, and comparing the classification results. The support vector machine model based on the TF-IDF algorithm performs the best, with a positive tendency accuracy, recall, and f1 of 72%, 69%, and 70%, respectively, for sentiment analysis of comment texts, and a negative tendency accuracy rate The recall rate and F1 were 71%, 73%, and 71%, respectively. Using the N-gram tool to construct and extract emotional word features, Bi gram and Tri gram, and using the word cloud tool based on the TF-IDF algorithm model to generate intuitive visual comment sentiment word cloud maps of positive tendency, neutral tendency, and negative tendency, respectively.