Question:1.
Predicting Prices of Used Cars. The file ToyotaCorolla.xls contains data on used Toyota Corollas for sale during late Summer of 2004 in The Netherlands. The dataset has 1,436 records containing details on 38 attributes, including Price, Age, Kilometers driven, Horsepower, and other specifications. The goal is to predict the price of a used Toyota Corolla based on its specifications (textbook reference - 11.3).
Data Preprocessing. Create dummy variables for Fuel Type. Eliminate the following columns: Id, Model, Met_Color, Color, Cylinders, Radio, and Radio_Cassette. Split the data into training (60%), validation (40%) datasets.
a. Use XLMiner's Neural Network routine to fit a model using XLMiner default values for the neural net parameters (including Normalizing the Input Data). Also, make sure to set the Error Tolerance = "0" before running the Neural Network algorithm. Record the RMS error for the training data and the validation data. Repeat the process, changing the number of epochs (and only change epochs) to 300, 3000, and 10,000.
i. What happens to the RMS error for the training data as the number of epochs increases?
ii. What happens to the RMS error for the validation data?
iii. Comment on the appropriate number of epochs for the model.
b. Conduct similar experiments and discuss the effect of changing the following in XLMiner (Note: the following experiments should be done independently of each other):
i. Increase the number of Hidden Layers in the network up to the maximum allowed in XLMiner
ii. Increase and Decrease the Gradient Descent Step Size (i.e., Learning Rate)
Attachment:- toyotacorolla.xls