^{1 } Another alternative to lasso is ridge regression which is expressed by Equation 5 except with an
L2 norm instead of an L1 norm for the regularization term. In Equation 5, the L1 norm penalizes the
absolute value of the coefficients, used by lasso; while in ridge regression, the L2 norm penalizes the
squares of all coefficients. Given this property, ridge regression is not as effective at penalizing
parameters to zero compared to lasso regression (Tibshirani, 1996). Therefore, lasso regression is
preferred for variable selection.