1: \begin{abstract}

2: Recent work in the field of signal processing has shown that the singular value decomposition of a matrix with entries in certain real algebras can be a powerful tool. %For example, the complex SVD has been used to separate dispersive waves, the polynomial SVD for convolutive mixtures, and the quaternion SVD for 3D polarised waves.

3: In this article we show how to generalise the QR decomposition and SVD to a wide class of real algebras, including all finite-dimensional semi-simple algebras, (twisted) group algebras and Clifford algebras. Two approaches are described for computing the QRD/SVD: one Jacobi method with a generalised Givens rotation, and one based on the Artin-Wedderburn theorem.

4: \end{abstract}