Consumer Preference Analysis and Rating Prediction Model in the Restaurant Industry Based on Restaurant Information and Consumer Reviews

The restaurant industry has increasingly relied on the development of the Internet and mobile apps in recent years. Diner tends to make decisions based on restaurant information and customer reviews on apps, while merchant also focuses on customer reviews to improve the quality of service and attract more customers. It is noteworthy these customer reviews and ratings show that consumers pay more attention to restaurant attributes such as environment, food variety, parking condition and the need for reservation. These obvious tendencies can either serve as positive factors for restaurants, or directly result in consumer dissatisfaction and negative reviews. However, many review apps limit consumer to rating on a scale of 1–5 with a difference of 0.5, which not only restricts customers’ ratings, but also affects the rating’s authenticity. Therefore, this paper will firstly explore the influence of different factors on consumer ratings by using regression analysis to summarize consumer’s preference and selection tendency. Secondly, it will compare the prediction results of consumer ratings by using regression model, decision tree model and random forest model. The result shows that random forest model can effectively predict consumer ratings, reduce rating errors while keeping MSE and R2 within a reasonable range, and reflect consumer’s real attitude towards restaurants with greater accuracy.