论文部分内容阅读
研究基于时间序列的感知QoS的云服务组合,将服务的QoS偏好随时间不断变化的过程纳入云服务组合的研究范围,将云服务组合建模成时间序列的相似度对比问题。分别用欧几里得距离和扩展Frobenius范数距离度量二维时间序列的相似度,继而用基于主成分分析的扩展Frobenius范数距离和欧几里得距离、Brute-Force等方法度量多维时间序列的相似度,通过实验对比验证扩展Frobenius范数距离度量相似度在时间和准确性上的优越性。
This paper studies the cloud service combination based on time series and perceived QoS. It takes the changing of service preference over time into the scope of cloud service portfolio, and models the cloud service portfolio into the comparison of time series similarity. The Euclidean distance and the extended Frobenius norm distance are used to measure the similarity of two-dimensional time series respectively. Then the extended Frobenius norm distance and Euclidean distance based on principal component analysis and the Brute-Force method are used to measure multi-dimensional time series The similarity of extended Frobenius norm distance metric similarity is proved by experiments to be superior in time and accuracy.