DefinePK

Abstract

The need for filtering authentic web services from the rest has become quite hectic seeing as the malicious ones seem tomore and more legitimate by the day. The loss of proprietary information to such web services is becoming increasinglycommon and as such the paper suggests an idea in which a third party authenticating web service evaluate web service providersand facilitates the user with credible list of providers based on its "Trust filter". The filter takes into account many factorsincluding the various customer preferences and prepares its itinerary trusted candidates in order to satisfy the customized needsof the individual. The idea is to mine user’s preferences from the requirements specification provided by the user and thepreferences are used to determine the weights of each Quality of Service QoS attribute. The local trust on a service for the user isderived by combining the trust on QoS attributes and the trust on user’s ratings. In order to classify the web services based on theuser's opinion the user must first be legitimate. The global trustworthiness of a service for the users group, the dishonest user areremoved based on the results of Local honesty evaluation process. The simulation results indicate that the model works well onpersonalized evaluation of trust, and it can effectively dilute the influence of malicious ratings.