MATRIX Course

Eigendecomposition steps

Go back

The steps for the eigendecomposition are

  • calculate $\lambda{I_n}-A_n$
  • solve the lambdas (eigenvalues) in $det(\lambda{I_n}-A_n) = 0$
    • you should have $n$ eigenvalues $\lambda_n$
    • the order does not matter
  • Now you got $D$, the eigenvalues are on the diagonal
  • for each eigenvalue
    • you can solve $A - \lambda_n * Id_n = 0$
    • you can write it as a system
\[ \begin{split}\left \{ \begin{array}{r c l} (a_1 - \lambda_n) x + b_1 y + ... = 0 \\ a_2 x + (b_2 - \lambda_n) y + ... = 0 \\ ... \end{array} \right .\end{split} \]
  • each result (the vector (x,y, ...)) is an eigenvector
  • by concatenating all of your eigenvectors, you get $P$
  • solve $P^{-1}$
  • check $P*D^1*P^{-1}=A$