Recurrent neural network for web services performance forecasting, ranking and regression testing

Accurate estimation of web services performance, which is critical to ensure the consumers satisfaction on web services is still a challenging task due to the dynamic, and personalized requirements of different individuals. Efficient estimation of web services performance can lead to a better ranking of web services. Regression testing is then performed on the ranked web services to ensure that existing functionality of the web services is not impacted through evolution in the web services. Soft computing techniques are highly resource consuming, and more complex for practitioners. Moreover, they show complex approximation with a low propagation, which can be improved by using the advanced deep neural networks. Previously proposed web services performance estimation and analysis have been never considered from the deep neural network. To address the problem of efficient estimation of web services performance, gated recurrent unit (GRU) has been proposed with the use of time slice quality of service (QoS) data of web services. The GRU model can analyze QoS values obtained from different sets of users in different timestamps. The proposed approach has been evaluated on the web services dataset and comparison indicates that the proposed approach shows the better prediction and estimation than the state of the art approaches.