1. For a data with 12 predictors and a response taking three values, we fit a logistic linear model. Answer the following questions. (a) We use 5-folds to get the following CV plot. The x-axis in the plot is in natural logarithmic scale. Find the optimal A. 10 10 10 10 10 10 10 10 8 8 8 7 7 4 4 4 3 2 2 2 2 1 1 0 4 Mulinomial Deviance -6 log (Lambda](h) Suppose the output of one lasso estimation is as follows. Write down the estimated model. > reg=glmnet(x,y,family='multinomial',lambda=0.05) > reg$a0 SO 0 *0.3969305 1 0.6031214 2 *0.2061909 > reg$beta $'0' 12 x 1 sparse Matrix of class "dgCMatrix\" 30 V1 *1.0690179 V2 V3 *O.6214503 V4 V5 V8 V7 V8 V9 V10 V11 V12 $1n 12 x 1 sparse Matrix of class "dgCMatrix" SO V1 V2 V3 V4 V5 V6 V7 V8 V9 0. 04550377 V10 V11 -0. 08720514 V12 $2' 12 x 1 sparse Matrix of class "dgCMatrix" SO V1 0.9308649 V2 V3 0. 9890976 V4 V5 V6 V7 V8 V9 V10 .V11 V12 0. 1438392 (c) Let A - co, explain what would happen in the lasso estimation