A Novel QoS Predication Approach Based on Regression Algorithm

Abstract

Recently, as an important way of software applications, web services technology developed rapidly. As many web services provide the same functionality, quality of service (QoS) has become an important criterion for service selection. However, the QoS value provided by service provider is being tested in an ideal runtime, regardless of the effect of the actual load and different service time. This paper uses web service runtime period and service load as criteria, considers QoS historical data, and proposes a novel real-time QoS predication approach based on dualistic linear regression algorithm, which could be used to select the optimal web services. Experiments demonstrate the efficiency of the approach.