Sparse reduced-rank regression for simultaneous rank and variable selection via manifold optimization

Abstract We consider the problem of constructing a reduced-rank regression model whose coefficient parameter is represented as singular value decomposition with sparse vectors. The traditional estimation procedure for often fails when true rank high. To overcome this issue, we develop an algorithm and variable selection via regularization manifold optimization, which enables us to obtain accurate even if Using regularization, can also select optimal rank. conduct Monte Carlo experiments real data analysis illustrate effectiveness our proposed method.