Publication

Abstract : In the present Internet era, with the seemingly insatiable growth in applications based on web services (WS), it is an arduous task for an user to select the foremost web service among a large number of competing services. Prediction of quality of service (QoS) ratings of WS helps to identify optimal WS. The ratings are appraised by an evaluation of different QoS factors. However, in the real world, missing values are often major problem in datasets of QoS factors, as they lead to imprecise prediction of QoS rating of a given web service. We present here a system designed for imputing the missing values in QoS datasets, by clustering, similarity checking and optimization techniques. The optimization of QoS values helps to obtain the more accurate dataset. Finally, QoS prediction of WS is calculated using imputed dataset. It helps to ascertain the quality and ranking of different WS. Experimental results show that the proposed method predicts web service QoS values more accurately and using these complete dataset an optimal web service is predicted/suggested to the user.