Text Mining Analysis of Online Reviews of Scenic Spots Based on LDA Topic Model—Taking Ctrip Nanxun Ancient Town Scenic Spot Review as an Example

互联网的快速发展给各大景区带来机遇的同时，也带来了挑战。很多人在互联网上进行景点评论搜索并根据景区评论选择目的地。通过分析游客的评论，旅游景区可以更好地找到景区存在的问题，以便提出相应的解决方案，从而提升游客的满意度和旅游体验。本文利用爬虫技术对携程旅行上南浔古镇景点的用户评论进行爬取，利用LDA主题模型、SnowNLP情感分析和TF-IDF特征提取技术，对爬取的评论数据进行了全面分析，深入探讨景区评论中消费者的满意度及情感分布，了解消费者对旅游景区的核心需求，为景区改善服务提供了数据支撑。The rapid development of the Internet has brought both opportunities and challenges to major scenic spots. Many people search scenic spot reviews on the Internet and choose destinations according to the reviews. By analyzing tourists’ comments, tourist attractions can better find problems existing in scenic spots, so as to propose corresponding solutions, thus improving tourists’ satisfaction and travel experience. This paper uses crawler technology to crawl the user reviews of scenic spots in Nanxun Ancient Town on Ctrip, and uses LDA theme model, SnowNLP sentiment analysis and TF-IDF feature extraction technology to conduct a comprehensive analysis of the crawled review data, deeply explore the satisfaction and emotional distribution of consumers in scenic spot reviews and understand the core needs of consumers for tourist attractions. It provides data support for the service improvement of scenic spots.