Modeling the Dynamic Trust of Online Service Providers Using HMM

Online trading takes place in a very complex environment full of uncertainty in which deceitful service providers or sellers may strategically change their behaviors to maximize their profits. The proliferation of deception cases makes it essential and challenging to model the dynamics of a service provider and predict the trustworthiness of the service provider in transactions. Recently, probabilistic trust models have been used to assist decision making in computing environments. Although the typical Hidden Markov Model (HMM) has been used to model a provider´s behavior dynamics, existing approaches focus only on the outcomes or ignore the hidden characteristics of the HMM model. In this paper, we model the dynamic trust of service providers concerning a forthcoming transaction in light of as much information as we can consider, including the static features, such as the provider´s reputation and item price, and the dynamic features, such as the latest profile changes of a service provider and price changes. Based on a service provider´s historical transactions, we predict the trustworthiness of the service provider in a forthcoming transaction. In addition, the Mutual Information theories and the Principle Component Analysis method are leveraged to eliminate redundant information and combine essential features to form lower dimensional feature vectors. Furthermore, by adopting Vector Quantization techniques, we apply the discrete HMM in a more powerful way, in which all the features extracted from both contextual information and the rating of each transaction are treated as observations of HMM. We evaluate our approach empirically in order to study its performance. The experiment results illustrate that our approach significantly outperforms the state-of-the-art probabilistic trust methods in accuracy in the cases with complex changes.