Web services reputation assessment using a hidden Markov model

Abstract

We present an approach for reputation assessment in service-oriented environments. We define key metrics to aggregate the feedbacks of different raters, for assessing a service provider's reputation. In situations where rater feedbacks are not readily available, we use a Hidden Markov Models (HMM) to predict the reputation of a service provider. HMMs have proven to be suitable in numerous research areas for modelling dynamic systems. We propose to emulate the success of such systems for evaluating service reputations to enable trust-based interactions with and amongst Web services. The experiment details included in this paper show the applicability of the proposed HMM-based reputation assessment model. © 2009 Springer-Verlag Berlin Heidelberg.