4. (35 marks) Variance and bias for Linear regression v.5. Ridge regression Fit the data with model i 4, : n | (313:; | 6;,1thBTB a: '13\": NW, 02}. Ridge regression parameter estimtates: [fifmge = Afidgei, and {fridge = % i=1 ' Least square parameter estimtates: 35'" = 37 stf, and [3113 A. For a new point In, calculate bias and variance for (3.) Linear regression (b) Ridge regression Where bias2 2 [g + [313:0 ]E(;i,}.g)]2 and variance 2 E[;i}g Egg]? (Note: Yon will see that compared with linear regression, the bias2 for ridge is larger, but the variance is smaller. At high dimensional setting, ridge regression ( and lasso) will be better because of the smaller variance.) Hint: You may directly use some derivations in the lecture. Var(A + B) = Var(A) + Var(B) + 2ch(A, B) VarfaA'I = QQthr'tAi. where c: is a scalar