Web Services QoS Measure Based on Subjective and Objective Weight

The previous Quality of Service (QoS) measure method for Web services are not accurate as they often focused on the ambiguity of user preferences but neglected its one-sidedness (i.e., user preferences cannot realize the data distribution characteristics of services set). This paper first points out that user preferences are not only ambiguous but also one-sided. QoS measure should be a co-action of both user preferences and services set, and therefore neglecting the data distribution characteristics of services set have weakened the accuracy of measure results. We present in this paper a novel QoS measure algorithm for Web services employing the subjective and objective weight. The subjective weight is used to quantify ambiguous user preferences and the objective weight is used to correct the one-sidedness of user preference. The two used weights guarantee the measure results can both conform to user preference and reflect the overall performance precisely. The validation of theoretical analysis and experiments based on the QWS real data sets confirmed the efficiency of proposed algorithm.