A K-Means Clustering Approach for PCA-Based Web Service QoS Prediction

Aiming at the problem of inaccurate Quality of service(QoS) Prediction results for Web service due to sparse matrix, a k-means clustering method is presented to predict Web service QoS through Principal Component Analysis(PCA). First, to obtain a QoS matrix without missing values, the original high-dimensional sparse matrix is filled by PCA dimensionality reduction. Second, the service users is clustered on the filled QoS matrix. Finally, the unknown service QoS values are predicted for the active users based on the clustering information. Comprehensive experiments are conducted on multiple real-world Web service QoS datasets. The experimental results indicate that our method achieves higher prediction accuracy in sparse matrix than other conventional methods.