Skip to content

Matrix Decomposition Selection Guide

Run the Selection Guide Fullscreen

Edit the MicroSim in the p5.js Editor

About This MicroSim

This interactive decision tree helps you choose the right matrix decomposition for your problem. Click on blue decision nodes to explore different paths, and green nodes show the recommended decomposition with reasoning.

Decision Summary

Goal Matrix Type Recommended
Solve Ax=b Symmetric positive definite Cholesky
Solve Ax=b Square, invertible LU with pivoting
Solve Ax=b Rectangular QR or SVD
Least squares Any overdetermined QR
Low-rank approx Any Truncated SVD
Eigenvalues needed Any SVD

How to Use

  1. Start at the top "Start" node
  2. Click on blue question nodes to explore paths
  3. Read the green answer nodes for recommendations
  4. Hover over nodes to see explanations

Learning Objectives

After using this MicroSim, students will be able to:

  • Match problems to appropriate decompositions
  • Understand why certain decompositions are preferred
  • Navigate the decision tree efficiently
  • Justify decomposition choices

Quick Reference

When to Use Each Decomposition

Decomposition Best For Complexity
Cholesky SPD systems, fastest O(n³/3)
LU Square systems, multiple b's O(2n³/3)
QR Least squares, rectangular O(2mn²)
SVD Low-rank, rank analysis O(mn²)

References

  • Chapter 7: Matrix Decompositions - Choosing the Right Decomposition