Quadratic Forms: Representation & Classification

Published on May 9, 2025 by Aman K Sahu

What is a Quadratic Form?

A quadratic form is an expression involving a quadratic polynomial of several variables. It is often represented using matrices for simplification and analysis.

For example, the expression:
Q(x) = 3x² + 4xy + 5y²
can be written in matrix form as:
Q(x) = [x y] × [[3, 2], [2, 5]] × [x y]ᵗ

💡 Notice how the cross term (4xy) becomes 2 in both symmetric off-diagonal positions: 2 + 2 = 4.

General Matrix Representation:

For a vector x = [x₁, x₂, ..., xₙ] and symmetric matrix A, the quadratic form is:
Q(x) = xᵗAx

Classification of Quadratic Forms

Based on the nature of Q(x) = xᵗAx, quadratic forms are classified as:

  • Positive Definite: Q(x) > 0 for all non-zero x.
    💡 All eigenvalues of A are positive.
  • Negative Definite: Q(x) < 0 for all non-zero x.
    💡 All eigenvalues of A are negative.
  • Positive Semi-Definite: Q(x) ≥ 0 for all x.
    💡 All eigenvalues are ≥ 0.
  • Negative Semi-Definite: Q(x) ≤ 0 for all x.
    💡 All eigenvalues are ≤ 0.
  • Indefinite: Q(x) takes both positive and negative values.
    💡 Eigenvalues have mixed signs.

Applications of Quadratic Forms

  • Optimization: In quadratic programming, the goal is to minimize or maximize a quadratic form subject to constraints.
  • Machine Learning: In algorithms like Support Vector Machines, kernels and margins use quadratic expressions.
  • Physics: In energy expressions like potential energy stored in springs.
  • Statistics: Used in analyzing variance and in least squares estimation.
Previous: Matrices & Operations
Next: Systems of Linear Equations