关键词: Web服务, 服务组合, 服务消亡, 特征提取, 机器学习, 泊松分解, 极限学习机

Abstract: Aiming at the problem that Web services deprecation had led the service compositions invoked this Web service to be invalidated,which impacted on the stability of service system.The topic features was extracted from services’ profiles,and a service deprecation prediction algorithm was proposed with Poisson Factorization (PF) and Extreme Learning Machines (ELM) methods,which combined with time factors.Experiments on a real-world dataset showed that the proposed model could effective predict Web services deprecation and provide suggestions on service reliabilities for service system managers and service composition developers.

Key words: Web service, service composition, service deprecation, feature extraction, machine learning, Poisson factorization, extreme learning machines