Feature Normalization

#Computers
In some machine learning algorithms, failing to normalize features results in that feature being over or underrepresented

$\displaystyle x_{\text{normalized}}= \frac{x-\bar{x}}{s_{d}}$

• Essentially a z-score but using sample metrics rather than population metrics
• $\displaystyle x$ is the original sample data
• $\displaystyle \bar{x}$ is the mean of $\displaystyle x$
• $\displaystyle s_{d}$ is the standard error in $\displaystyle x$
• This is the equivalent of the StandardScaler() in sklearn

$\displaystyle x_{\text{normalized}}=\frac{x-x_{\text{max}}}{x_{\text{max}}-x_{\text{min}}}$