Discovering Important Services Based on Weighted K-Core Decomposition

With the development of service-oriented architecture, the number of services is expanding rapidly. Important services usually have high quality, and they can be recommended to users if the users do not give any keyword. However, how to discover the important services is still a problem facing many people. In this article, the authors propose a novel approach to discover important services based on service networks. First, their approach uses service networks to abstract services and the relations between them. Second, the authors employ the weighted k-core decomposition approach in the field of complex networks to partition the service network into a layered structure and calculate the weighted coreness value of each service node. Finally, services will be ranked according to their weighted coreness values in a descending order. The top-ranked services are the important ones the authors' approach recommends. Experimental results on a real-world data set crawled from ProgrammableWeb validate the effectiveness of their approach.