Like Diagonalization, but you perform it on an orthogonal matrix, which makes the process easier as follows:

Where , i.e. is orthogonal It is otherwise the same as performing diagonalization on an arbitrary matrix