基于Word2vec-BiLSTM-Attention的餐饮评论情感分析方法研究OA

To fully grasp the sentiment tendencies of users in catering reviews and assist merchants in precisely optimizing services,a sentiment analysis model integrating Word2vec,bidirectional long short-term memory network and Attention mechanism is proposed.This method uses Word2vec to train word vectors to represent semantic information,captures context features through BiLSTM,and introduces an Attention mechanism to focus on sentiment keywords.The experiment is based on the catering review data of Meituan-Dianping,and LSTM,BiLSTM and BiLSTM-Attention are selected as the baseline models for comparative analysis.The results show that the accuracy rate of the constructed Word2vec-BiLSTM-Attention model is 90.75%,and the F1 value reaches 91.13%.Its comprehensive performance is superior to that of the LSTM and BiLSTM baseline models,verifying the effectiveness of this combined model in the task of sentiment analysis of catering reviews.