Polynomial Regression
Linear regression on polynomial feature expansions for curved relationships
Polynomial Regression generates interaction and power terms from the original features, then fits a linear model on the expanded set. This allows it to capture nonlinear relationships while remaining interpretable.
When to use:
- Known nonlinear relationships (quadratic, cubic) between features and target
- Small feature sets where polynomial expansion is manageable
- When a smooth curve fits the data better than a line
Input: Tabular data with the feature columns defined during training Output: Continuous predicted value
Model Settings (set during training, used at inference)
Degree (default: 2)
Polynomial degree. 2 adds squared terms and pairwise interactions; 3 adds cubic terms. Higher degrees risk overfitting.
Interaction Only (default: false) If true, only interaction features are produced — no polynomial powers (e.g., x² is excluded).
Include Bias (default: true) Whether to include a bias column of ones.
Fit Intercept (default: true) Whether the linear model includes an intercept.
Inference Settings
No dedicated inference-time settings. The same polynomial transformation applied during training is applied to inputs before prediction.