Dynamic web service recommendations based on model based collaborative filtering methods and Genetic Algorithm

As cloud computing has emerged as new computing paradigm, more and more web services has been provided on the Internet, thereby how to select a qualified service is becoming a key issue. Because of service users locations and the conditions of network, the nonfunctional Quality of Service(QoS) attributes has to be considered for web service selection. The contribution of this paper is three fold for dynamic web service recommendations. At first, this paper proposes a new approach which fuses both Slope One and Item based Clustering Collaborative Filtering(CF) methods for predicting the QoSvalues of web services and recommends the best services which meet the user requirements. Secondly, a refined K-Means algorithm(RKMA) is proposed which improves the prediction accuracy when compared tostandard K-Means(KMA) algorithm. Finally, this work also proves that quality of the clusters can be improved by using Genetic Algorithm(GA).