Research of Web Service Recommendation Using Bayesian Network Reasoning

摘要

How to recommend the atomic and a set of services with correlations to meet users' functional and non-functional requests is a key problem to be solved in the era of services computing. On the basis of organizing service clusters with different functions using the three-stage Bayesian network structure learning method. It uses the parameter learning method to obtain the conditional probability table (CPT) of all the nodes. The Bayesian network reasoning method (Gibbs Sampling) is used to recommend a set of service types that are interested to users. Finally, it selects a set of services in the specific service clusters to meet users' functional and QoS requirements. The case study and experiments are used to explain and validate the effectiveness of the proposed method.