Online word-of-mouth content mining is of great significance to product, service improvement and demand prediction of online marketing enterprises. However, most studies have focused on the identification of the sentiment tendency of online word-of-mouth, and lack of text content mining for online word-of-mouth, especially negative word-of-mouth. This paper introduces deep learning into online word-of-mouth sentiment tendency analysis and negative word-of-mouth word attribute feature extraction, and builds an online word-of-mouth sentiment tendency analysis and attribute extraction model based on LSTM deep learning algorithm. The model was trained and tested through online word-of-mouth data of a fashion apparel e-commerce company. The results show that the LSTM model has a good effect on sentiment analysis and negative word-of-mouth attribute feature extraction. Through comparative experiments, it is shown that the model has a better effect than the traditional machine learning methods (SVM, Naive Bayes) in the analysis of sentiment tendency.
Yang, Jiamei; Li, Li; and Lin, Yulan, "Research on Online Word-of-mouth Sentiment Analysis and Attribute Extraction Based on Deep Learning" (2020). WHICEB 2020 Proceedings. 32.