标准化/规范化通常是针对每个属性单独完成的,就像您现在所做的那样。但为了回归,请咨询
https://stats.stackexchange.com/questions/29781/when-conducting-multiple-regression-when-should-you-center-your-predictor-varia/111997#111997
据我所知有两点,
min-max normalization
z-score normalization
公式,其中A是数据集中的属性 -
# Min-Max Normalization (Final values are in between 0 and 1) v_ = (v - min(A)) / (max(A) - min(A)) # Z - Score Normalization (Final values have a mean of 0 and SD of 1) v_ = (v - mean(A)) / (standard_deviation(A))
实现完全取决于编程语言。例如,在R中,您可以使用函数扫描将其标准化为一行