Menu

Feature Scaling

Feature Scaling

Machine learning algorithms often perform better when numerical features are on a similar scale.

In the credit card dataset, most variables have already been transformed using Principal Component Analysis (PCA). However, columns such as Amount and Time may still require scaling.

Feature scaling helps:

  • Improve model convergence
  • Prevent large values from dominating smaller ones
  • Increase training efficiency

Common scaling techniques include:

  • Standardization
  • Min-Max Scaling

In this project, feature scaling ensures that all input variables contribute appropriately during model training.