Multivariate-local forecasting model of principal component analysis for cloud computing
面向云计算的主成分分析多变量局域预测模型

To solve the problems that cloud computing system can not get enough forecasting information from the univariate load sequence, this paer proposed the multivariate-local forecasting model based on principal component analysis, and applied it in the forecasting of the underlying resource for cloud computing. Using of principal component analysis method to consider the relationship between the underlying resources, it determined the embedding dimension of multivariate phase space and combine with the local forecasting method. The simulation results show that the multivariate-local forecasting model based on principal component analysis can provide more precise of forecasting than the univariate-local forecasting model. Thus, the multivariate-local forecasting model is demonstrated to be efficient for the forecasting of the underlying resource for cloud computing.