Determinants, Row-Echelon Form

Row-Echelon Form

Any square matrix m can be transformed into a diagonal matrix, i.e. a matrix that is zero everywhere except the main diagonal, in a manner that preserves the determinant of the matrix. The matrix has been converted into "row-echelon form". First perform gaussian Elimination, giving an upper triangular matrix having the same determinant. Then perform back substitution to clear out everything above the main diagonal. The result is a diagonal matrix having the same determinant.

Actually gaussian elimination might multiply the determinant by -1, if we are forced to swap rows. For most applications the sign of the determinant doesn't matter, or can be inferred. If it does matter, keep track of the swaps.