Why using Feature Scaling ?

Check this data set for example:

• The feature level has a much lower numerical value compare to salary even though it is also very significant to the output of the prediction.
• For situation like this we apply feature scaling to enhance the precision and accuracy of the result
• When to NOT apply feature scaling: (click to expand)

The math behind basic Feature Scaling:

Standard Scaling Formula:

Here is the formula for standard scaling (z-score normalization):

$$ z = \frac{x - \mu}{\sigma} $$

where:

• $z$ is the standardized value,
• $x$ is the original value,
• $\mu$ is the mean of the original values,
• $\sigma$ is the standard deviation of the original values.

Why this formula ?

Standard scaling, also known as z-score normalization, is a technique used to standardize the range of independent variables or features of data. The goal is to transform the data to have a mean of 0 and a standard deviation of 1. This is particularly important for machine learning algorithms like Support Vector Regression (SVR), which are sensitive to the scale of input features. By applying this transformation, the machine ensure that each feature contributes equally to the distance calculations, improving model performance and convergence.