WEB Service Selection Approach Based On QoS Prediction In Mobile Internet Environments

Currently,as the number of Web services increasing fast,more and more Web services share the same functionality.It is difficult for users to select Web services with better QoS(quality of service)when needed.Meanwhile,users would get different QoS of a same Web service because of the different network condition they have.For the above reasons,many researchers start to study the method of selecting Web services based on QoS prediction and have made some remarkable achievements.But in the environment of mobile internet,most of these methods would result in large prediction deviation because of the volatility.To solve the problems mentioned above,a Web service selection method based on QoS prediction is proposed in this paper.We propose the following optimizations:1)We use preprocessing method for data preprocessing.The impact of volatility of history QoS data to the following predictions is efficiently reduced.And the network volatility feature is preserved to a certain degree.2)An improved PCC similarity measurement method and an effective similarity attenuation method are combined together to enhance the accuracy of similarity calculation.3)With the method of screen used interval and the calculation of similarity weight,the prediction accuracy increases and the impact of volatility is reduced.4)As for Web service selecting,we use user's tendency to judge.Considering the different preferences of a user to different QoS attributes,we try to make the best Web service selection.Experimental results show that our approach can reduce the impact of volatility and improve the quality of service selection.