Credibility Assessment And Recommended Methods Web Services

Web-based service compositions have the following advantages:rich functional, cross-platform, convenience and so on. Because of its advantages, Web-based service compositions are changing the traditional software development model. As the development of web service composition, the web service type is more and more rich in Web, the number of service is becoming more and more huge too, hundreds of Web services on the one hand, provides people with more choices, On the other hand, they have no ideal when faced with choice, people have to consider whether or not the service safety, suitability and other issues. Therefore, an effective service recommendation system is indispensable; it can be recommended to the users of the users are faced with the problem. In such cases, this thesis researches trustworthy services recommendation, main purpose is to find a feasible service recommendations to recommend suitable service to users.In this thesis, the main works are trustworthy attribute level assessment, trustworthy service recommendations and attributes conflict during the trustworthy service recommendations. In trustworthy attribute level assessment, on the basis of analyzing and summarizing the previous study, integration and improvement of previous studies to make it fit to trustworthy attribute level assessment. In this respect, the main work of this thesis have:1. Putting forward a idea of constructing a trustworthy attribute tree,2.Analyzing, summarizing and improve the previous study, made the study to adapt the new environment of trustworthy attribute level assessment.3. When multiple experts node trust value is calculated considering the weight of each expert.In the above study, we study the trustworthy service recommendations, and the reason of studying trustworthy attribute level assessment is to provide data for trustworthy service recommendations. Because of commerce recommendation and services recommendation have some similar, we first studied the commerce recommendation and services recommendation, and summarized the similarities and differences of the two, and found they are different in rating the items:In e-commerce recommendation user rating of the product is the only one, however, in services recommendation ser rating of the product would change In the different environment. At the same time, in service recommendations, after the services composition, services’ trustworthy attribute are not a simple line relation, some attribute will promote each other, such as safety and reliability, some attribute will weaken each other, such as safety and ease of use. Therefore, after the services composition each trustworthy attribute will influence each other, In order to make the recommendation service can meet user demand after the service composition, we must solve the problem of attribute conflict. Based on these, we proposed a trustworthy service recommendations method, the method belongs to the collaborative filtering recommendation method, This method come from improvements on the old method. In this part the main work of this paper have:1.In-depth analysis and summary the differences and similarities of commerce recommendation and services recommendation.2.Put forward a recommendations method.3.Solving the problem of attributes conflict during service recommendations.At the end, based on the above service recommendations method, we have designed a service recommendation system, simulation and experiment analysis to the recommendations method, the experiments show that the method can meet the needs of users for trustworthy service choice.