Normalization: normalization, which generally operates on data before training, eliminating the mapping of dimensions to [0, 1]
Standardization: normalize the data, so that the mean 1 variance is 0. It is also the processing of the data
Regularization: Regularization, the penalty term for LOSS prevents overfitting.
Note the difference between normalization and normalization :
Normalization treats different feature information equally
The purpose of normalizing the scaling of different feature dimensions is to make features comparable between different metrics. At the same time, the distribution of the original data is not changed.
