Question 2: Case Two–German Credit
Dataset: CreditRisk.xls
In this task, your goal is to build a decision tree model to identify inputs or predictors that differentiate risky customers from others (based on patterns pertaining to previous customers) which can then be used for predicting new risky customers.
The dataset to be used for the task is CreditRisk.xls. The Credit dataset (based on a German dataset from ftp.ics.uci.edu/pub/machine-learning-databases/statlog/ renamed CreditRisk and altered*) has 1,000 records and 30 variables pertaining to past and current customers who have borrowed from a bank for various reasons. The data contains customer-related information such as financial standing, reason for loan, employment, demographic information. Each application was rated as “good credit” (700 cases) or “bad credit” (300 cases) based on the institution’s past experience.
*The original dataset had a number of categorical variables, some of which have been transformed into a series of binary variables so that they can be handled appropriately by XLMiner.
Your Task:
Partition the data into training and validation datasets. Build a classification tree model to learn the characteristic of the problem. Test its performance on the validation sets. Prepare a report that identifies the decision tree model and training parameters, as well as the resulting performance on the validation sets, together with overall discussion of limitations and suggestion on the performance improvement.
Your report should include:
(1) Problem domain, data pre-processing, and model building
This includes the problem issues to be addressed, the rationales of each of the data mining steps taken e.g. data examination and pre-processing, model building including justification of the techniques and explanation of the approaches taken, make reference to your XLMiner Excel file where necessary
(2) Results interpretation and overall discussion
This includes interpretation of the…