Programming Homework Help

(Based on Problem 3) Consider the data on used cars (ToyotaCorolla.csv) with 1436 records and details on 38 attributes, including Price, Age, KM, HP, and other specifications. The goal is to predict the price of a used Toyota Corolla based on its specifications.

1.Convert any categorical predictors to dummies.

2.Scale the numerical predictors and the outcome variable to a 0–1 scale. You can scale all of the columns with a single command using the predict function and the preProcess function from the caret package:

my.scaled.df <- predict(preProcess(my.df, method = “range”), my.df)

3.Partition the data into a training set (60% of the rows) and a validation set.

4.Fit a neural network model to the data. Use a single hidden layer with 2 nodes. As mentioned above, use the predictors Age_08_04, KM, Fuel_Type, HP, Automatic, Doors, Quarterly_Tax, Mfr_Guarantee, Guarantee_Period, Airco, Automatic_airco, CD_Player, Powered_Windows, Sport_Model, and Tow_Bar.

5.Report the RMSE on the training data and the validation data.

6.Repeat the process, changing the number of hidden layers and nodes to (1) a single layer with 5 nodes, and (2) two layers with 5 nodes in each layer. Report the RMSE on the training data and the validation data each time.

7.What happened to the RMSE on the training data as the number of layers and nodes increased?

8.What happened to the RMSE on the validation data?

9.Of the three neural nets that you tried, which would you recommend be used?

I will sent the dataset later