Personalized Online-to-Offline (O2O) Service Recommendation Based on a Novel Frequent Service-Set Network

Rapid development of information technology allows consumers to search and purchase services or products online, and then redeem them in offline stores. This emerging e-commerce model, the online-to-offline (O2O) service platform, has been catching significantly increasing attention. However, the numerous cyber O2O services lead to difficulties for customers to make appropriate decisions. In this paper, we propose a personalized O2O service recommendation method based on a novel frequent service-set network (FSSN). This method integrates the service network attribute based on FSSN and service essential attributes, which include popularity and commonality of service itself, and then generates a service recommendation list, which largely addresses the problem of sparse matrix. The experimental results show that FSSN-based recommendation has better performance than similarity-based recommendation methods and state-of-the-art matrix methods. Meanwhile, we find the relative importance degree of service network attribute and service essential attributes, and find the optimized combination of popularity performance and commonality performance to form service essential attributes with different matrix densities.