So I want to get sparse coefficients after training a LASSO model but I want the sparsity to be maintained on each column of the coefficients.
Suppose after training sklearn.linear_model.Lasso with 5 input dimension and 3 output dimension I get the following coeff _ :
array([[ 1.87666825 , 0. , -1.37516633, 0.57491936, -0.65120491],
[ 0. , -0.5901364 , -0.85214947, 0.03498916, -0.81805396],
[ -0.89515717, -0.04372684, -0.22986802, 0. , -0.56785277]])
Is it possible to get sparsity on a particular column? For example:
array([[ 0. , 1.87666825, -1.37516633, 0.57491936, -0.65120491],
[ 0. , -0.5901364 , -0.85214947, 0.03498916, -0.81805396],
[ 0. , -0.04372684, -0.22986802, -0.89515717, -0.56785277]])