Canonical low rank approximations (LRA) for regression in tensor decomposed format for Matlab. Contributed to automated hyperparameter optimization routines, abstrated and decoupled LRA algorithm from basis set selection and intermediate optimization routines.