2005 International Conference on Neural Networks and Brain
DOI: 10.1109/icnnb.2005.1615001
Full text: Download
A novel L1 significant vector (SV) regression algorithm is proposed in the paper. The proposed regularized L1 SV algorithm finds the significant vectors in a successive greedy process. The performance of the proposed algorithm is comparable to the OLS algorithm while it saves a lot of time complexities in implementing orthogonalization needed in the OLS algorithm