上QQ阅读APP看书，第一时间看更新
Data scaling
The values of each feature in a dataset can vary between random values. So, sometimes it is important to scale them so that this becomes a level playing field. Through this statistical procedure, it's possible to compare identical variables belonging to different distributions and different variables.
Remember, it is good practice to rescale data before training a machine learning algorithm. With rescaling, data units are eliminated, allowing you to easily compare data from different locations.