Decision Tree Programming Assignment

Learning Objectives

After completing this assignment, students should be able to:

Implement a decision tree for both classification and regression
Describe how tree depth relates to underfitting and overfitting

Part 1: Implementation

Complete the following stubbed-out decision tree classifier so that all public methods and attributes correspond to the provided docstring comments:

decision_tree.py

This module contains a method for drawing 2D classification trees:

draw_tree.py

You may find it useful for confirming that your finished classifier is working correctly.

Part 2: Analysis

For this part of the assignment you will apply your decision tree to the problem of determining a respondent’s age based on their answers to an online quiz. The quiz consists of 30 Yes/No questions like “Have you ever broken a bone?”, “Have you ever been on the radio or television?”, etc.

This may seem like a trivial exercise, but keep in mind that companies spend a great deal of time and money attempting to infer demographic information based on seemingly innocuous data revealed online.

For this part of the assignment you should submit a Jupyter notebook that satisfies the following requirements:

It should illustrate hyper-parameter tuning using cross validation.
It should illustrate model evaluation by determining the classification error on the test set.
It should include a 4x4 confusion matrix illustrating classification results on the test data.
It should include a discussion of which attributes were used by your decision tree.

Each step of the notebook must be accompanied by text explaining the point of the provided Python code and discussing the results.

Data Set

The following files contain the training and test data you must use for your analysis:

The class labels are integers in the range 0-3 where 0: 0-18 years, 1: 19-24, 2: 25-34, 3: 35+. The attributes are 0 or 1 where 1 corresponds to an answer of “Yes” and 0 corresponds to an answer of “No”. The thirty questions are:

Have you ever…

broken a bone	had your cell phone ring at an embarrassing moment
been on the radio or television	gotten stitches
gotten a speeding ticket	paid for a stranger’s meal
gone skydiving	had a tooth pulled out
been to Europe	held a butterfly in your hand
driven a boat	been pooped on by a bird
been on water skis	been stung by a jellyfish
been on a motorcycle	had braces
been horseback riding	had contacts or glasses
ridden on a bull	been in a major natural disaster
ridden on a sheep	fallen down the stairs
stayed awake for an entire night	faked being sick
been alpine skiing	sung in the shower
ran a marathon	gotten locked out of your car or house
regifted a gift	been shown on a JumboTron at a sporting event

This data is a processed version of the Have You Ever data set collected by Benjamin Soyka and distributed under the CC BY-SA 4.0 license.

Partners

This assignment may be completed individually or in pairs. My expectation for pairs is that both members are actively involved, and take full responsibility for, all aspects of the project. In other words, I expect that you are sitting down together to work, not that you are splitting up tasks to be completed separately.

If you intend to work with a partner, you must inform me no later than Tuesday 9/14.

Grading

Grades will be calculated according to the following distribution.

Readability/Style 10%

Your code should follow PEP8 conventions. It should be well documented and well organized.
Reference Tests 60%
- decision_tree_classifier_tests.py
- decision_tree_regressor_tests.py
Efficiency 10%

Our main concern with this assignment is clarity and correctness. That said, your implementation must be efficient enough execute the provided testing code in no more than a second or two. This means you should avoid Python loops where possible.
Part 2 Submission 20%

Details TBA.