基于用户偏好的可信QoS服务选择

发布时间：2018-05-20 03:24

本文选题：用户偏好 + 用户评价　；参考：《南京邮电大学》2016年硕士论文

【摘要】：随着互联网和电子商务技术的不断发展,越来越多的用户和服务参与到服务计算中,为用户从海量服务中准确选择出真正符合用户真实需求的服务变得日益重要。在服务选择过程中,可信QoS的有效获取作为一个重要环节正逐渐吸引着越来越多研究人员的注意。当前大多可信QoS的求取研究并未结合需要选择服务的目标用户的具体偏好开展,往往难以为该用户提供符合其真实感观的选择依据。针对以上问题,本文主要基于目标用户偏好进行可信QoS服务选择方法的研究,首先引入服务多属性评分和目标用户偏好,根据历史用户评分同主流评分的偏离度来衡量其相对于目标用户的可信任度,如果偏离度超过一定的阈值,该历史用户就会被记录一次,当被记录的次数达到另一阈值时,将该历史用户剔除出去,如此迭代执行,最终剩余用户即是可提供可靠参考评价的可信用户。然后,利用可信用户的服务评价求取服务的基准评价向量,基于服务的基准评价向量同可信用户评价向量进行相似度计算以确定候选服务的综合评价向量。接着求取可信用户对候选服务的多属性的满意程度,再结合候选服务的综合评价向量、目标用户偏好,确定候选服务的最终评分并选取评分值最高的服务返回给目标用户。本文通过应用实例演示了方法的可用性,其有效性也通过仿真实验得到了验证。
[Abstract]:With the continuous development of Internet and e-commerce technology, more and more users and services participate in service computing. It is increasingly important for users to accurately select services that really meet the real needs of users. In the process of service selection, the effective acquisition of trusted QoS, as an important link, is attracting more and more researchers' attention. At present, most of the research on trusted QoS has not been carried out in combination with the specific preferences of the target users who need to select the services, and it is often difficult to provide the users with a basis for their selection in accordance with their true sense of feeling. In order to solve the above problems, this paper mainly studies the method of trusted QoS service selection based on target user preference. First, we introduce service multi-attribute score and target user preference. The degree of trust relative to the target user is measured by the bias between the historical user score and the mainstream score. If the deviation exceeds a certain threshold, the historical user is recorded once, when the number of times recorded reaches another threshold. If the historical user is removed and executed iteratively, the end residual user will be a trusted user who can provide reliable reference evaluation. Then, the benchmark evaluation vector of the service is obtained by using the service evaluation of trusted users, and the similarity calculation between the benchmark evaluation vector based on service and the evaluation vector of trusted user is carried out to determine the comprehensive evaluation vector of candidate service. Then the degree of satisfaction of trusted users to the multiple attributes of candidate services is obtained. Then combining the comprehensive evaluation vector of candidate services and the preference of target users the final score of candidate services is determined and the service with the highest score is selected to be returned to the target users. The availability of the method is demonstrated by an application example, and its validity is verified by simulation experiments.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.09

【相似文献】