Live Streaming Product Recommendation Algorithm Based On Sentiment Analysis

Currently,live streaming e-commerce has become a new economic development model.In this process,users’ textual expressions,barrage comments,and voice in the live broadcast room are all their emotional expressions.Correctly identifying users’ emotional needs would greatly benefit live streaming e-commerce.A personalized recommendation system based on sentiment analysis needs to address how to accurately determine users’ emotional tendencies,how to effectively integrate different emotional expressions,and how to accurately analyze Chinese information.Traditional methods such as Chinese text segmentation technology and quantitative models for sentiment analysis face challenges of difficult feature extraction,immature sentiment quantification models,and difficulty in combining different emotional expressions.This article addresses the aforementioned issues and its contents are as follows:(1)An analysis of the basic structure of the LSTM model leads to the proposal of an improved LSTM model that enhances the accuracy of Chinese text segmentation and the extraction of emotional features.(2)A sentiment quantification processing model is constructed to improve the accuracy of determining users’ emotional tendencies during the process of live streaming e-commerce.The extensive barrage texts published by users during the process will be analyzed for their emotional words,negation words,and adverbs,which will then undergo sentiment quantification processing to determine their reasonable emotional values.(3)Combining the above models,an improved collaborative filtering recommendation algorithm is proposed to enhance the accuracy of live streaming e-commerce recommendations.The paper carries out experimental verifications on the proposed models,and the results show that the improved LSTM model improves the accuracy of Chinese text segmentation,while the quantitative sentiment model used in the collaborative filtering recommendation algorithm effectively combines comment information and ratings in the barrage texts to improve the accuracy of product recommendations.