4. This problem asks you to discuss both ridge regression and the lasso. (a) (10 points) A problem with ridge regression is that ordinary inference procedures (standard errors, p-values, etc.) are not applicable since exact distributional properties of the ridge estimates of the 3 coefficients are not known. Suggest a reasonable) procedure for approximating the standard errors of the regression coefficients, so that one could still draw conclusions about whether or not a particular coefficient is significantly different from zero when doing ridge regression. (b) (10 points) The procedure in part (a) could be used if we wanted to do ridge regression and still be able to select a subset of predictors predictors to include in our linear model. On the other hand, the procedure in (a) wouldn't be necessary if we wanted to choose predictors. Why is that? 4. This problem asks you to discuss both ridge regression and the lasso. (a) (10 points) A problem with ridge regression is that ordinary inference procedures (standard errors, p-values, etc.) are not applicable since exact distributional properties of the ridge estimates of the 3 coefficients are not known. Suggest a reasonable) procedure for approximating the standard errors of the regression coefficients, so that one could still draw conclusions about whether or not a particular coefficient is significantly different from zero when doing ridge regression. (b) (10 points) The procedure in part (a) could be used if we wanted to do ridge regression and still be able to select a subset of predictors predictors to include in our linear model. On the other hand, the procedure in (a) wouldn't be necessary if we wanted to choose predictors. Why is that