Research On Predicted Methods Of Service Quality And User Satisfaction Under Cloud Platform

With the development of Cloud Computing,the number of cloud service increases rapidly,and more and more diverse resources start to rely on access to the cloud.Different cloud services may have different Quality of Service(QoS)and user satisfaction,even though they have the same function.This research tries to figure out an effective way to recommend the best cloud service to the customer with the considering of QoS and user satisfaction.According to the real demand for cloud service,this research analyses the relationship between QoS and user satisfaction.The main contents are as follows:(1)The advanced processing of multiple-valued QoS ought to aim at the time series characteristics of QoS and the characteristics of multiple-valued QoS evaluation.To build could model in different time periods,a measurement method of similarity of integrated cloud model with shape and distance has been proposed,which can accurately identify the adjacent users of potential users.Besides,it would consider the potential users' demand during diverse periods to determine the objective weight of the time period by using the FAHP method.This collaborative QoS prediction method based on the analysis of the periodic changes of QoS is able to offer precise predicted results.(2)The second-order hidden markov model has been proposed to build a QoS satisfaction predict model,in allusion to user satisfaction.This model considers the effect of the previous two-time states on the current state,predicts the QoS satisfaction,and then recommends the resulting service QoS to the user with a higher ability to meet their requirement.This method can effectively improve the prediction accuracy and contribute to selecting the cloud service with higher user satisfaction.