Normalization of Data
Terminology
Normalization
Normalization refers to transforming given data to a value between $0$ and $1$. It is commonly achieved through a transformation $T$ using the data’s maximum value $x_{\text{max}}$ and minimum value $x_{\text{min}}$. $$ T (x) = {{ x - x_{\text{min}}} \over { x_{\text{max}} - x_{\text{min}}}} $$ Sometimes, it’s simply called Scaling.
Description
Note that in linear algebra, the term normalized vector refers to a vector that has been orthonormalized, which is a distinct concept from normalization.
Regularization, although similar in wording and context, refers to techniques related to loss in training rather than altering the data itself, marking a clear distinction.
See Also
- Standardization: Typically refers to the process in statistics of adjusting data to a mean of $0$ and a variance of $1$.
- Normalization: Usually describes the process of positioning data within a specific interval.
- Regularization: Often refers to the processes in machine learning to prevent overfitting.