Predicting the influence of group buying on the restaurant´s popularity by online reviews

Group buying, which provides considerable discounts for various services and products, is a novel marketing strategy that has been very popular all over the world in recent years. Customers, merchants and websites all can gain benefits from successful group buying campaigns. However, group buying is not always effective, merchants have to seriously consider that whether they are under appropriate business conditions and adopt proper group buying strategies. Different from the perspective of customers and websites held in most existing work, we focus on the impact of group buying on the operation performance from the aspect of merchants, especially restaurants. Based on online user reviews, we try to predict a restaurant´s popularity, which is an important measurement of its operation performance, during and after a group buying campaign. We consider this problem as a time series prediction problem. Under the framework of a typical intervention analysis model, we propose a new approach to handle this problem. A comprehensive experimental study on the real-world dataset has been conducted to demonstrate that our proposed methods are effective and outperform existing methods in predicting a restaurant´s popularity.