Diversified Quality Centric Service Recommendation

In the past decade, a large number and variety of services have been published on the Internet, e.g., public services, web services, cloud services, etc., offering customers a wide range of choices for fulfilling their demands. This has made it a challenge for customers to select from many mutually-substitutable competing services, especially when they are offered with differentiated quality. Thus, service recommendation that helps customers find the right services has become of paramount research and practical importance. The correlations in customers' preferences for different quality dimensions make service recommendation an even more complicated problem. It has not been properly addressed by existing approaches. This paper proposes DQCSR (diversified quality centric service recommendation), an approach that finds diversified services that are representative in different quality dimensions with respect to customer's quality preferences. The results of experiments conducted on a real-world dataset demonstrate the representativeness and diversity in the recommendation results produced by DQCSR.