Basis iteration for reward based dimensionality reduction | IEEE Conference Publication | IEEE Xplore