This assignment needs to be completed in SPSS… For each question, there needs to be screen shots for each step of each question with completed questions. Second, previously I've gotten many wrong answers… please I need correct answers to the questions!

The purpose of this assignment is to perform neural netsclassification, interpret the results, and analyze whether or not theinformation generated can be used to address a specific businessproblem.

For this assignment, you will use the "Credit Card DefaultsOversampled" data set from the Topic Materials. Note that this data setincludes the same information as the Topic 5 assignment data set,however, the default cases have been randomly oversampled to balance thedefaulters vs. the nondefaulters. This is one method to increase theaccuracy of a rare outcome (i.e., default).

You are an analyst for a credit card company. Management wants toknow if there are any early signs that indicate whether customers willdefault on their credit cards. If these indicators can be identified,then more scrutiny can be placed on customer transactions in an effortto avoid losses. For this project, management is interested in veryaccurate predictions. Management would like to receive a list of rankedvariables related to default, however, understanding the specificrelationships between the predictors and the likelihood of default issecondary.

Question 1: Partition the data to create a training data set (70%) and test data set (30%).

Question 2: Build a neural network usingthe Multilayer Perceptron method with the training data and "Default" asthe target. Make sure to output predictor importance. Include the"Default Neural Network" output when submitting the answer.

1. What is the ranking of the variables? Include the "Predictor Importance" output when submitting the answer.
2. What is the accuracy of the model when using the training and testdata? Include the "Misclassification Table" outputs when submitting theanswer.

Question 3: Rerun the model but only use the top six variables.

1. What is the accuracy of the model when using the training and testdata? Include the "Misclassification Table" outputs when submitting theanswer.
2. Did the accuracy improve? Explain why.
3. Produce a ROC curve when using the training and test data andinterpret the results of the ROC curves. What do the curve and the45-degree line represent?

Question 4: Explain how the selected modelcan be used to solve the business problem. What are the advantages anddisadvantages when building a neural net model vs. the simplerclassification model like the one completed in the Topic 5 assignment?

General Requirements:

Submit the answers to Questions 1-4 as a Word document.