Abstract: Taking Yantai Apple as the research object, the Python was used to crawl the user’s comment data under relevant products as the corpus, the natural language processing technology of Baidu AI platform was introduced, and the trained model was used to analyze the comment text. LDA theme classification model was built to extract text themes and keywords, and conduct text mining and quantitative analysis for online comments. The results showed that in the comment text, the positive evaluation for Yantai Apple accounted for 96.1%, neutral evaluation for 0.3%, and negative evaluation for 3.6%. According to the theme consistency index, the comment text content could be clustered into five themes： variety identification, service perception, product quality, marketing perception and logistics quality. Consumers’ overall satisfaction with Yantai Apple was high. Strengthening the variety identification of consumers was an important improvement direction in the future, and suggestions for further improving consumers’ satisfaction were put forward.

Key words: online review, consumers’ satisfaction of agricultural products, Yantai Apple, sentiment classification, LDA