Scaling

Standardization and normalization are the two terms for scaling techniques used in the industry. Both these techniques ensure that the numerical features used in the model are weighted equally in their representation. Most of the time people use standardization and normalization interchangeably. Though both of them are scaling techniques, there is a thin line of difference between the two.

Standardization assumes the data to be normally distributed. It rescales the data to mean as zero and standard deviation as one. Normalization is a scaling technique that assumes no prior distribution of the data. In this technique, the numerical data is rescaled to a fixed range either: 0 to 1, -1 to +1, and so on.

The following are a few widely ...

Get Hands-On Automated Machine Learning now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.