Abstract
A good location is critical for the sales performance of a hotel or a restaurant. This study proposes a machine learning-based model for location selection that focuses on the estimation of sales potential of a prospective site and helps to overcome the lack of historical data for the prospective site and the subjective criteria used in conventional models. The proposed model involves three major steps. First, we use an attribute selection algorithm to identify the key factors that contribute to the profitability of a specific location. Second, we evaluate the similarity between the candidate site and the existing stores by using an improved grey comprehensive evaluation method. Finally, we use a kernel regression model to predict the sales potential of the candidate site. A case study of a well-known international restaurant chain is used to illustrate the application of the proposed data-driven model. The results indicate that our proposed model helps to accurately select the most profitable locations.