Decision Tree Courswork for COMP70050 Introduction to Machine Learning

Introduction

This repository contains the code for the coursework for COMP70050 Introduction to Machine Learning. The coursework is to implement a decision tree algorithm and to test it on two datasets. The first dataset is a clean dataset and the second dataset is a noisy dataset. The code is written in Python 3.7.3.

Dependencies

The code requires the following dependencies:

• Python 3.7.3
• numpy 1.16.4
• pandas 0.24.2
• matplotlib 3.1.0
• scikit-learn 0.21.2

Contributors

• Mahanoor Syed
• Brendon Ferra
• Harry Phillips
• Ameen Izhac

Running code on default datasets

To run the code on the clean and noisy datasets please run: python3 main.py This will create trees that are unpruned, pruned and pruned with a depth limit for both datasets.

Running code on custom datasets

Please move your dataset into the wifi_db folder.

To generate an unpruned tree please run: python3 classify_dataset.py <custom_dataset.txt> <k>

To generate a pruned tree please run: python3 prune_dataset.py <custom_dataset.txt> <k>

To generate a pruned tree with a depth limit please run: python3 prune_dataset_limit.py <custom_dataset.txt> <k> <limit>

Report

The report for this coursework can be found here.

Score

We scored 100% for this coursework.