WSWalker: A Random Walk Method for QoS-Aware Web Service Recommendation

Recently, collaborative filtering has been applied to QoS-aware Web service recommendation. However, it cannot make recommendations for users that have invoked only a very small number of services because of data sparsity. In addition, these methods do not know how confident they are in their recommendations. Based on the fact that QoS values of web services are usually subject to the locations of users, a few works assume that the additional knowledge of users´ locations can be used to better deal with the data sparsity issue, since a user only needs to know the users near to him/her. On the other hand, the sparsity of user-service invocations forces the location-aware method to consider the QoS experiences of users not near enough, which may decrease its precision. In order to find a good trade-off between coverage and precision, we propose a random walk method combining location-aware and collaborative filtering method for web service recommendation. The random walk method allows us to define and to measure the confidence of a recommendation. To evaluate the performance of our proposed method, we conduct a set of comprehensive experiments using a real-world web service dataset, and compared the method with existing collaborative filtering methods.