Quadratic Regression Calculator

Fit a second-degree polynomial (y = ax² + bx + c) to your data using the least squares method. Enter x and y data points to find the best-fit quadratic curve.

QUADRATIC EQUATION
--
Coefficient a
--
Coefficient b
--
Coefficient c
--
R² (Fit Quality)
--

What Is Quadratic Regression?

Quadratic regression fits a second-degree polynomial curve (parabola) to a set of data points. Unlike linear regression which fits a straight line, quadratic regression can model curved relationships where the rate of change itself changes. The model has the form y = ax² + bx + c.

This technique is widely used in physics (projectile motion), economics (diminishing returns), biology (growth curves), and engineering. If your data shows curvature that a straight line cannot capture, quadratic regression may provide a significantly better fit.

The Least Squares Method

The coefficients a, b, and c are found by solving a system of three normal equations derived from minimizing the sum of squared residuals:

y = ax² + bx + c
Minimize: Σ(yi - axi² - bxi - c)²

The R² (coefficient of determination) measures how well the quadratic model fits the data, ranging from 0 (no fit) to 1 (perfect fit).

When to Use Quadratic vs Linear

FeatureLinear RegressionQuadratic Regression
Modely = mx + by = ax² + bx + c
ShapeStraight lineParabola
Min data points23
Best forConstant rate of changeAccelerating/decelerating trends

Frequently Asked Questions

How many data points do I need?

You need a minimum of 3 data points to fit a quadratic model (since there are 3 coefficients). However, for reliable results with meaningful R² values, at least 5-10 data points are recommended. More data generally yields a more reliable model.

What does R² tell me?

R² (coefficient of determination) indicates how well the quadratic model explains the variation in your data. An R² of 0.95 means 95% of the variation is explained by the model. Values above 0.9 generally indicate an excellent fit. If R² is low, consider whether the relationship is truly quadratic or requires a different model.