Regression trees model a relationship by providing constant values on rectangles that partition the input space. O True O False Cost-complexity pruning of a regression tree is used to decrease the chance of overfitting. O True O False Classification trees are examples are nonparametric models. True O False The Patient Rule Induction Method (PRIM) results in a regression tree via a different approach. True Falsee. [3.75 pts each] Given the Following shorthand canonical SUP expression: f(x,y,z] =Zm(e,2, 3. :5, T} a. List the complete until table for f{x,y,z}. h. Express f{x,y,z} in Fully expanded SUP canonical Form. c. Express f{x,y,z] in shorthand P05 canonical Form. :1. Express f{x,y,z] in Fully expanded PBS canonical Form. 7- [3-75 PIS each] Given the following shorthand canonical PDS expression: flames]: H M[1,3,4,7,11,12,14} a. List the complete IJ'utJI table for fight}, as}. h. Express f{p,q,r,s] in Fully expanded PBS canonical Form. c. Express f{p,q,r,s] in shorthand SOP canonical Form. d. Express f{p,q,r,s] in Fully expanded 5GP canonical Form. The following equation is estimated, where you is also an explanatory variable and the following results are obtained: y, = 2.7 +0.4x, + 0.9y,-1 (0.4) (0.06) n=200, R? =0.98, and DW=1.9 a. What does the Durbin Watson statistic test? What assumptions are used to calculate the Durbin- Watson statistic? For a DW of 1.9, what does this imply about the regression equation? b. Irrespective of your answer in part (a), how would you correct for AR(1) in this model?Multiple Choice The rpart() function in R, A fits a regression tree using recursive partitioning B fits multiple regressions in local neighborhood C fits a regression tree using bootstrap aggregation D fits a non linear regression in local neighborhood SUBMIT I Multiple Choice In regression trees, A the predictor space will not be splitted into sub regions B a single predictive model explains the entire predictor space C the predictor space is recursively partitioned, and the average response value of each partition is used as the model estimate D the predictor space is recursively partitioned, and will be fitted a linear model for each partition SUBMIT