✦ScienexySCIENCE KNOWLEDGE LIBRARY
Science reference · SCI-CS-003-073

Polynomial Regression

Polynomial regression models nonlinear relationships by adding polynomial powers of input variables to a regression model.

ŷ = β₀ + β₁x + β₂x² + ... + βₙxⁿ

Explanation

Polynomial regression is an extension of linear regression in which input variables are transformed into polynomial terms such as squared or cubed values. These additional terms allow the model to represent curved relationships between inputs and a continuous target. For example, a quadratic model includes a squared input term, while a cubic model also includes a third-power term. The coefficients are usually estimated using Ordinary Least Squares. Although the resulting curve may be nonlinear with respect to the original input, the model remains linear in its coefficients. The polynomial degree controls the flexibility of the curve. A degree that is too low may underfit the data, while a very high degree may overfit noise and perform poorly on unseen observations. Regularization and validation can help control model complexity.

Variables

SymbolMeaning

ŷ

Predicted target value

x

Original input variable

β₀

Intercept

βᵢ

Coefficient of the polynomial term

n

Polynomial degree

Example or application

A model predicting crop growth from temperature may discover a curved relationship. A quadratic regression model can represent growth that increases with temperature up to an optimum and then decreases.

Text

Polynomial regression allows a regression model to fit curved patterns while remaining linear in its coefficients.

Key Points

  • Polynomial regression adds powers of input variables.

  • It can represent curved relationships.

  • The degree determines the model's flexibility.

  • It remains linear in its coefficients.

  • High-degree models can overfit.

  • Validation and regularization can help control complexity.

Applications

  • Growth curve modeling

  • Temperature-response prediction

  • Population trend analysis

  • Engineering measurements

  • Scientific data fitting

Learning Outcomes

  • Define polynomial regression.

  • Explain how polynomial features create curved predictions.

  • Distinguish polynomial degree from the number of input features.

  • Understand the risks of overfitting.

Practice Questions

  • What is polynomial regression?

  • How does a quadratic model differ from a linear model?

  • Why is polynomial regression still considered linear in its coefficients?

  • What happens when the polynomial degree becomes excessively high?

Related Concepts

  • Linear Regression

  • Polynomial Features

  • Underfitting

  • Overfitting

  • Regularization

Common Mistakes

  • Assuming polynomial regression is nonlinear in its coefficients.

  • Choosing a very high degree without validation.

  • Ignoring numerical instability caused by large polynomial powers.

  • Confusing a curved fitted relationship with a nonlinear optimization problem.

Sources and verification

Sources are shown as provided in the project data. Not every source has yet been independently verified. Please report errors or missing citations.

  • James, G., Witten, D., Hastie, T., Tibshirani, R., and Taylor, J. (2023). An Introduction to Statistical Learning.
  • Bishop, C. M. (2006). Pattern Recognition and Machine Learning.

Suggest a correction or source · More in Artificial Intelligence & Machine Learning