a$)$ First use a Type node to specify the role of each attribute and double check the measurement of attributes based on data description if it is not correctly defined.

b$)$ Train a neural networks model using Neural Net node.

c$)$ Generate a visual representation of the neural networks model. Report the top $5$ attributes ranked by the model. $($Note: this ranking could be different than what reported by the decision tree model in Part $($b$)).$

d$)$ You extracted some rules after building decision tree in part $($b$).$ Can you extract rules when you build a neural networks model? Why?

e$)$ Evaluate and analyze the model's performance metrics. Report Confusion Matrix, Accuracy for both testing and training sets. If there are $2$ price bins, report AUC.

f$)$ Is the model overfitted? Why? If yes, address this problem. Hint: One option is to reduce the number of attributes by selecting only the top few attributes $($explain how many top attributes you would like to select?$).$ Another option is to set the parameters to opt for lower number of nodes in the hidden layer $($if you take this approach, report on the parameters setting$).$

$($d$)$ Models Comparison

$1-$ Compare the performance of the best decision tree you trained with the best neural networks in terms of Accuracy. $($If there are $2$ price bins, report AUC$).$