Refer to the previous exercise for a description of the data set. Create a regression tree model for predicting house prices (Price). Select the best-pruned tree for scoring and display the full-grown, best-pruned, and minimum error trees. 

a. Display the best-pruned tree. How many leaf nodes are in the best-pruned tree? What are the predictor variable and split value of the root node of the best-pruned tree? 

b. What are the RMSE and MAD of the best-pruned tree on the test data? On average, does the regression tree over-or under-predict prices of houses? Is the regression tree model effective in predicting prices of houses?

c. Score the two new houses on the market in the Houses_ Score worksheet using the best-pruned tree. What are their predicted prices according to your model?

Data from Exercises 36

Melissa Hill is a real estate agent in Berkeley, California. She wants to build a predictive model that can help her price a house more accurately. Melissa has compiled a data set in the House_Data worksheet that contains the information about the houses sold in the past year. The data set contains the following variables: number of bedrooms (BM), number of bathrooms (Bath), square footage of the property (SQFT), lot size (Lot_Size), type of property (Type), age of the property (Age), and price sold (Price). A portion of the data set is shown in the accompanying table. Build a default regression tree to predict house prices (Price). Display the regression tree.

Transcribed Image Text:

BM Bath SQFT Lot_Size 4 2 1520 2 1251 E 4314 3 E 4 E 4 Type 4050 Single Family 3200 Single Family 1 B 7056 Single Family Age 110 112 1 78 Price 500000 775000 1 2635000