This exercise will guide you through the process of creating a Neural Network using Pytorch for data about a car

market. You can use much of the code from the class lecture on PyTorch. For this problem, you need to

understand and study the model covered in the class.

Read the data

$1.$ You will do this problem in a Jupyter Notebook and submit a Jupyter Notebook file, Problem$2.$ipynb

$2.$ You have been given the data $$cars$.$csv$$

$3.$ You have been given the file $$Problem$2.$ipyb$$

$4.$ If necessary, run the Kernel$->$Restart Kernel and Clear All Outputs command.

Build a Model

$1.$ Create a NN model

a$.$ There should be three hidden layers

b$.$ The outer layer $($h$1)$ should have $9$ neurons

c$.$ The middle layer $($h$2)$ should have $8$ neurons

d$.$ The last layer should $($h$3)$ should have $7$ neurons

e$.$ They should be fully connected

f$.$ You can select any activation function

g$.$ The input feature layer size depends on the number of features you will

select from the data after co$-$relation analysis

h$.$ The output layer depends on the number of classes in your data

Clean and prepare the data $($training and testing$)$

$1.$ Clean the data. You will be building a NN to classify a car for carbody $($Convertable$,$

hatchback,sedan$)$

$2.$ Determine the co$-$relation of features with respect to the carbody

$3.$ Select the top five features to represent a car object$/$data

$4.$ Train and slit the data for training and testing

$5.$ Select a loss function

$6.$ Select an optimizer

$7.$ Select epochs $($ min $100)$

$8.$ Training and test data

$9.$ Test the model and print the loss performance