基于主分量分析的多维时间序列特征选择方法研究

研究了多维时间序列的特征选择问题。首先，讨论了主分量分析的原理；然后，将提出的斜率趋势算法和核函数分类算法应用于多维时间序列的特征选择过程中，并探讨了两类方法的优缺点；最后，采用两种算法对目标距离像多维时间序列进行了仿真实验。实验结果表明了所应用方法的可行性和有效性.

Research on Character Choosing of Multivariate Time Series Based on the Principle Component Analysis

The character choosing of multivariate time series was studied in this paper. The principle component analysis principle was introduced. Two approaches based on the slope threshold and kernel function classification wcre used for character choosing of multivariate time series. The advantages and disadvantages of the two methods were discussed. The results of experiment with target showed that the approaches were effective and feasible.