Linear Algebra & Matrices

Gram–Schmidt Calculator

Turn independent columns into an exact orthogonal basis and an approximate orthonormal basis.

Linear Algebra & Matrices

Enter matrix values

Exact structural results; labelled numerical approximations where required
  1. 1EnterProvide the known values
  2. 2CalculateResults update automatically
  3. 3VerifyReview the details and units
Try an example
Vectors as columns of A

Use signed finite decimals. Fill every visible cell. Fractions, commas, and scientific notation are not accepted.

Linear algebra result

Enter valid values to see the result.

Your entries are calculated in this browser and are not submitted to 365CALCS.COM.

Feedback

Understand the subject

What does Gram–Schmidt do?

Gram–Schmidt turns independent vectors into perpendicular directions by subtracting each earlier projection.

Use it to construct an orthogonal or orthonormal basis for the same span.

The relationship

The defining relationship

See the structure

What the calculation is doing

Worked interpretation

Read the result in context

uₖ=vₖ−Σprojᵤⱼ(vₖ)

Interpret with care

Important boundary

Near-dependent numerical inputs can be sensitive; the page labels approximate normalization where needed.

Use the Matrix Calculator to experiment with small exact matrix operations before applying a more specialized route.

Quick guide

How to use this calculator

  1. Set the matrix dimensions required by the task.
  2. Fill every visible entry and any operation-specific control.
  3. Read the result type, factors, classification, and disclosed limitations.

Calculation method

Apply the stated linear-algebra contract

uₖ = vₖ − Σⱼ<ₖ projᵤⱼ(vₖ)

Undefined, singular, dimensionally incompatible, dependent, or unsupported inputs receive an explicit message instead of a fabricated numeric result.

Worked example

Example interpretation

Normalization is approximate when a vector length is irrational.

uₖ = vₖ − Σⱼ<ₖ projᵤⱼ(vₖ)

Supported inputs

Precision and limits

Matrix size

Input matrices are limited to 1–6 rows and columns; operation-specific square, rank, or compatibility rules still apply.

Entry precision

Enter signed finite decimals with at most 30 digits and 15 decimal places.

Numerical policy

Results use real double-precision arithmetic, a 1e-10 structural tolerance where stated, bounded iterations, and labelled approximations.