Fit multinomial logistic regression with a penalty on the nuclear norm of the estimated regression coefficient matrix, using proximal gradient descent.
conda install r::r-npmr
