Dynamic Trust Evaluation Mechanism Based On Perception Of Affective Intensity

In recent years, with the rise of the Internet, the applications of open computing system(Agent system, P2 P network, Ad Hoc network etc.) and electronic commerce(eBay, Amazon etc.) have developed rapidly. However, its dynamic characteristics such as openness, anonymity and deceptive have made the security problem more and more prominent, and the security problem has become a key factor restricting the development of network. Therefore, it’s important to construct a reasonable and effective trust evaluation mechanism. The existing trust evaluation mechanism have some shortage, such as they generally didn’t fully consider the uncertainty problem of trust, ignored the description of negative and hesitation information, did not take into account the inconsistency between text comments and ratings, and can’t meet the requirements of users on the accuracy of evaluation etc..Therefore, a dynamic trust evaluation mechanism based on the perception of affection intensity. In this model, the sentiment analysis technology is integrated into the trust management, the expressions of trust is redefined by emotional intensity vector, and the dynamic trust relationship between entities are evaluated by the trust property data, so to get the real reasonable assessment results.Firstly, collect review information of user terminal, the review information is a combination of evaluation of multidimensional trust attributes. We extract the trust property and the corresponding emotion words by object extraction technique, and the emotional intensity in each polarity(positive, negative and neutral) of sentiment words are calculated according to the sentiment dictionary, it realized the correspondence between trust attribute evaluation and emotional intensity vector. So the model make the description of trust more reasonable and complete, take over the uncertainty of trust and eliminate the problem of inconsistent between the real evaluation and grade evaluation.Secondly, based on the quantitative trust evaluation data, we designed the trust computation algorithm based on the unit of trust property data. The algorithm calculates the trust value of service entity by combining the history record and recommended data. Considering the dynamic characteristics that trust will decay with time, we set the time weighted by exponential distribution method, these processes can achieve the effective containment of malicious behavior. Then, combined with the score function and accuracy function to compare the overall trust value, and gave the service selection strategy.Finally, based on the above theory, this paper design and implement a trust evaluation mechanism based on the perception of affective intensity, and the performance of trust evaluation mechanism is studied by simulation experiment. The experimental results show that the proposed model can effectively describe the trust relationship between nodes, significantly identify and eliminate various malicious attacks, and improve the accuracy of trust evaluation.