Free Cholesky Decomposition Calculator

Matrix Decomposition Basics

In linear algebra, a matrix decomposition (or factorization) rewrites a given matrix as the product of two or more simpler matrices, analogous to factoring an integer into prime factors. Many decomposition methods exist — among the most widely used are the LU decomposition, QR decomposition, singular value decomposition (SVD), and the Cholesky decomposition. Each method is designed for specific types of matrices and applications. The Cholesky decomposition, in particular, provides a fast and numerically stable way to work with a certain class of matrices, and is often implemented in scientific computing, optimization, and machine learning.

Defining the Cholesky Decomposition

The Cholesky decomposition (also called the Cholesky factorization) factorizes a real, symmetric, and positive‑definite matrix $A$ into the product of a lower triangular matrix $L$ and its transpose $L^{T}$:

$$A = L L^{T}.$$

For this factorization to exist, $A$ must satisfy three conditions:

• $A$ must be symmetric, i.e. $A^{T} = A$.
• Hence $A$ must be square.
• $A$ must be positive definite (all its eigenvalues are positive, or equivalently $z^{T} A z > 0$ for every nonzero vector $z$).

When these requirements are met, $L$ is unique and its diagonal entries are real. The lower triangular structure of $L$ — zeros above the main diagonal — makes many subsequent computations (e.g., solving linear systems) much simpler.

Computing Cholesky Factors

There is no single closed‑form formula for the Cholesky decomposition, but an algorithmic process exists: entries of $L$ are computed from the top‑left corner to the bottom‑right. For a general $n\times n$ matrix, the following recurrence yields the elements of $L$.

Diagonal entries ( $i = j$ ):

$$L_{ii} = \sqrt{ A_{ii} - \sum_{k=1}^{i-1} L_{ik}^{2} }$$

Off‑diagonal entries ( $i > j$ ):

$$L_{ij} = \frac{1}{L_{jj}} \left( A_{ij} - \sum_{k=1}^{j-1} L_{ik} L_{jk} \right)$$

These formulas show that each element depends on previously computed entries. The process can be applied to matrices of any size; the online Cholesky factorization calculator automates it for the most common sizes: 2×2, 3×3, and 4×4.

Example: 2×2 Cholesky Decomposition

For a 2×2 matrix $A$, the recurrence simplifies to:

$$L_{11} = \sqrt{A_{11}}, \quad L_{21} = \frac{A_{21}}{L_{11}}, \quad L_{22} = \sqrt{A_{22} - L_{21}^{2}}.$$

Take $A = \begin{pmatrix} 4 & 2 \\ 2 & 5 \end{pmatrix}$. Then

$$L_{11} = \sqrt{4}=2,\quad L_{21} = \frac{2}{2}=1,\quad L_{22} = \sqrt{5-1}=2,$$

so

$$L = \begin{pmatrix} 2 & 0 \\ 1 & 2 \end{pmatrix}.$$

Multiplying $L L^{T}$ recovers the original $A$, confirming the decomposition.

Example: 3×3 Cholesky Decomposition

Now consider the 3×3 symmetric positive‑definite matrix

$$A = \begin{pmatrix} 4 & -2 & 2\\ -2 & 5 & 1\\ 2 & 1 & 6 \end{pmatrix}.$$

The entries of $L$ are obtained in a fixed order:

1. $L_{11} = \sqrt{4} = 2$
2. $L_{21} = \dfrac{-2}{2} = -1,\quad L_{31} = \dfrac{2}{2} = 1$
3. $L_{22} = \sqrt{5 - (-1)^{2}} = \sqrt{4} = 2$
4. $L_{32} = \dfrac{1 - [(-1)(1)]}{2} = \dfrac{1+1}{2} = 1$
5. $L_{33} = \sqrt{6 - (1^{2}+1^{2})} = \sqrt{4} = 2$

Hence

$$L = \begin{pmatrix} 2 & 0 & 0\\ -1 & 2 & 0\\ 1 & 1 & 2 \end{pmatrix}.$$

One can verify that $L L^{T}$ equals the original $A$. The same pattern extends to 4×4 matrices and beyond.

Using the Online Cholesky Factorization Calculator

The Cholesky decomposition calculator on this page makes the factorization effortless. Simply:

1. Select the matrix size — 2×2, 3×3, or 4×4.
2. Fill in the matrix elements in the provided grid.
3. Click the button to compute.

The tool instantly returns the lower triangular matrix $L$. If the input matrix is not symmetric or not positive definite, the calculator will alert you. It also supports displaying the result with rational or decimal numbers, making it ideal for homework checks, engineering analysis, or data science tasks.

Why the Cholesky Decomposition Matters

Because $L$ is lower triangular, the factorization $A = L L^{T}$ drastically speeds up many linear algebra operations. Solving a system $A x = b$ becomes a two‑step substitution process (forward followed by backward substitution) that is much faster than general Gaussian elimination. The Cholesky method also requires less memory and is more numerically stable for symmetric positive‑definite matrices. These advantages make it the method of choice in applications like Monte Carlo simulations, Kalman filtering, finite‑element analysis, and portfolio optimization.