中文说明：用于信号特征提取、信号消噪，正确率可以达到98％，实现了对10个数字音的识别程序可以实现模式识别领域的数据的分类及回归，一种噪声辅助数据分析方法,Relief计算分类权重。
English Description：
Used for signal feature extraction, signal denoising, the correct rate can reach 98%, of 10 digital audio identification procedures can be achieved in the field of pattern recognition of data classification and regression, a noise assisted data analysis method, classification weight calculation of relief.