Sentiment Analysis and Fine-Grained Feature Mining of Hotel Reviews Based on BERT-BiLSTM
TL;DR
This paper proposes a hybrid neural network model combining BERT and BiLSTM, and verifies it on the hotel review data set, and provides a complete analysis framework from emotion classification to problem positioning for the hotel industry, which has important theoretical value and practical significance.
Abstract
With the rapid development of online tourism platform, hotel reviews published by users on the platform have become important data resources that affect consumer decision-making and improve service quality. For the sentiment analysis of Chinese hotel reviews, this paper proposes a hybrid neural network model combining BERT and BiLSTM, and verifies it on the hotel review data set. At the same time, combining LDA topic modeling and TF-IDF keyword extraction technology, fine-grained feature mining is carried out for negative comments. The experimental results show that the BERT-BiLSTM hybrid model achieves 93.10% accuracy and 93.19% F1 score on the test set, which is significantly better than the single BiLSTM and BERT model. LDA revealed that the negative comments mainly focused on the two core themes of front desk service and check-in process, hardware facilities and comprehensive experience. TF-IDF keywords further quantified the user’s focus. This study provides a complete analysis framework from emotion classification to problem positioning for the hotel industry, which has important theoretical value and practical significance.