通过主成分分析可以对混合物量测矩阵进行svd分解，截取特征值大的变量，可以滤掉一些无关信息，使计算量更小(Through the principal component analysis of the mixture can be measured matrix svd decomposition, interception of large eigenvalues of variables, you can filter out some irrelevant information, so that the calculation of smaller)