Abstract: In the “fuzzy and autonomy” cloud computing environment,the service category is numerous and the quality is uneven,so it is difficult for the users to carry out a trusted service selection.On the basis of the user interaction experience and the real interpersonal interaction model,a comprehensive trust evaluation model based on collaborative re-commendation was proposed.In the model,two kinds of dynamic factors,the time decay and the weights,are introduced and a multivariate hybrid collaborative recommendation algorithm is designed to achieve effective collaboration among users,which can help users select the trusted cloud services.In order to verify the feasibility of the model,a distributed prototype system was designed,and the simulation experiments were carried out on the user satisfaction and service quality of the model.Simulation results show that the model can faster increase the average service satisfaction and more effectively inhibit malicious service,and with the growth of the number of interactions,the service selection quality will also continue to improve.

Key words: Trust evaluation,Service selection,Trust model,Collaborative recommendation,Cloud computing