CSCF: A Mashup Service Recommendation Approach based on Content Similarity and Collaborative Filtering

Abstract

Lightweight Mashup service become very prevalent now since there are lots of advantages for them, such as easy use, short development time, and strong scalability. It is a challenge problem how to recommend user-interested, high-quality Mashup services to user with the rapid development of more and more Mashup service. In this paper, we propose CSCF (a Mashup service recommendation approach based on Content Similarity and Collaborative Filtering). CSCF firstly computes the content similarity between user history records and Mashup services and gets user interest value. Secondly, according to Mashup QoS(Quality of Service) invocation records of user, user similarity model and service similarity model are designed, and then get the QoS prediction value of active user to target service by using collaborative filtering. Finally, combining user interest value and QoS predictive value of Mashup service, CSCF ranks and recommends Mashup services to user. The experiments are performed with real Mashup services dataset, and the results of experiments show that CSCF can effectively recommends Mashup services to user with well-interesting, high-quality, better prediction precision.