Additionally this method is computationally much less intensive by holding the dimensions of the feature matrix in the presence of long records.

• 通过保持目标特征矩阵结构不变,解决了现仃方法在长采样数据条件下计算量过大的问题。