Sentiment Analysis of Tourist Reviews on Historical and Cultural Districts Based on Tourism Reviews: A Case Study of Fujian Province

China's five thousand years of rich history have nurtured many historical and cultural districts with profound cultural heritage. In recent years, these districts have become popular tourist destinations for both domestic and international travelers. This paper, using Fujian Province as a case study, conducts sentiment analysis and LDA topic modeling on online reviews to explore the factors influencing tourist satisfaction with historical and cultural districts and proposes suggestions for improving tourist satisfaction. Additionally, addressing the current issue of low accuracy in sentiment classification of online reviews, this paper introduces an improved model based on RoBERTa_wwm word embeddings and Bidirectional Gated Recurrent Unit (BiGRU), incorporating an attention mechanism to highlight the emotional words in the text that have a greater impact on classification results, thereby enhancing the accuracy of sentiment classification. Experimental results demonstrate that the model performs exceptionally well across various datasets. Moreover, LDA topic modeling reveals that tourist reviews are primarily focused on local characteristics, with particular attention given to aspects such as the history and culture, local specialties, snacks, and architecture of historical and cultural districts.