Machine Learning · Chapter 10 of 40
Polynomial Regression
POLYNOMIAL REGRESSION fits a curve by adding powers of the feature (x², x³, ...) as extra columns.
Still linear in the coefficients — just uses non-linear features.
Example 1 (python)
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
pf = PolynomialFeatures(degree=2)
X2 = pf.fit_transform(X)
LinearRegression().fit(X2, y)Add x² features.
Example 2 (python)
# degree=10 -> overfit; keep it lowHigher degree = more risk of overfitting.
Key points
- Fits curves using x², x³, ...
- Uses linear regression under the hood.
- Higher degree = risk of overfitting.
- Great for non-linear-but-smooth data.
💡 Note: Combine with a Pipeline so you can grid-search the best degree.