Predicting Quality of Service (QoS) Parameters using Extreme Learning Machines with Various Kernel Methods

Web services which are language and platform independent self-contained web-based distributed application components represented by their interfaces can have different Quality of Service (QoS) characteristics such as performance, reliability and scalability. One of the major objectives of a web service provider and implementer is to be able to estimate and improve the QoS parameters of their web service as its clients application are dependent on the overall quality of the service. We hypothesize that the QoS parameters have a correlation with several source code metrics and hence can be estimated by analyzing the source code. We investigate the predictive power of 37 different software metrics (Chidamber and Kemerer, Harry M. Sneed, Baski & Misra) to estimate 15 QoS attributes. We develop QoS prediction models using Extreme Learning Machines (ELM) with various kernel methods. Since the performance of the classifiers depends on the software metrics that are used to build the prediction model, we also examine two different feature selection techniques i.e., Principal Component Analysis (PCA), and Rough Set Analysis (RSA) for dimensionality reduction and removing irrelevant features. The performance of QoS prediction models are compared using three different types of performance parameters i.e., MAE, MMRE, RMSE. Our experimental results demonstrate that the model developed by extreme learning machine with RBF kernel achieves better results as compared to the other models in terms of the predictive accuracy.