please solve this

Problem 4 (3 points) In a regression problem for (Y, X1, X2, X3), assume that the true model is Y = f(X1, X2, X3) + E, where f(X1, X2, X3) = X2 + X3+ X3, and e is a random noise with mean 0. We can fit the following three models in R, (MI) Y ~ X1 + X2 + X3, (M2) Y ~ X2 + X3+ X3, (M3) Y ~ X1 + X2+ X3+ X? + X3+X3, Under each model, we can construct an estimator of the regression function usually denoted by f. Please rank the three models in terms of (a) bias of f, (b) variance of f, and (c) expected test MSE (for example, MI