Sparse learning has been proven to be a powerful technique
in supervised feature selection, which allows to
embed feature selection into the classification (or regression)
problem. In recent years, increasing attention
has been on applying spare lear