CSCI 479 -- Machine Learning
Fall 2026 - Assignment 2
Submit deadline: 12:30pm, 16 October 2026, Friday

Objectives:

• Practice building a decision tree model
• Practice using the decision tree model
• Practice documenting the process of building and using the decision tree model

References:

• Lab 3
• Assignment 1
• Information Based Learning

Problem Description:

The original German Credit Dataset used in Assignment 1 consists 20 descriptive attributes of various types. The Modified German Credit Dataset removed quite some descriptive attributes and only have the following handful of descriptive attributes left: "purpose", "employment", "housing", "job", "foreign_worker".

This reduction may compromize the quality of the predictive model trained from the data, but it makes building a decision tree attainable even partially manually.

Additionally, the dataset is split to two subsets. In the worksheet "Training Dataset", 800 data items are selected to train a decision tree model. And in the worksheet "Testing Dataset", 200 data items are saved to test the decision tree model built from the training dataset.

Your Tasks:

You are asked to perform and document the process of building a decision tree from the training dataset provided, and applying the built decision tree to predict the credit risk of each data item in the testing dataset.

The process and your written document should include at least the following steps:

• First, perform a quick data exploration of both the dataset and make sure the data are in suitable format for decision trees. You can re-use some of your Assignment 1 results.
• Train a decision tree using the given training dataset. present the built tree in a human readable way. (You may borrow the idea of how an expanded directory is displayed in the side bar of the file explore on Windows.)
  For each leaf node in your decision tree, determine and display its associated class label and briefly justify why it is assigned this label.
  No matter whether you used a machine learning platform or your own program, explain and justify your choice of the first descriptive attribute placed in the root node of your decision tree.
• Apply your built decision tree on each of the data items in the testing dataset. Explain how the prediction is achieved. Properly record the prediction results.
• Compare the actual credit risk of the data items in the testing dataset with your model's predictions. Explain how the comparison is done. Present the comparison result in a confusion matrix.
  Coupled with the cost matrix, calculate the cost score of your decision tree model.

How to submit:

To submit your assignment, login to your VIU Learn account, find the CSCI 479 course page, click on the "Assessment" drop-down menu, click on the "Assignments" item, then click on the folder named "A2". Then you can click on the "Add a File" button to browse and upload your report and any other associated files.

Your report file should be in one of the following formats: pdf, docx, md, and txt, preferably pdf. If you need to submit your own program, only the source code files should be included.

If you have any problem with submitting on the VIU Learn, please email me to let me know and attach your solution files with your email.