Fundamentals of Machine Learning Project

Overview

This project applies various machine learning techniques to predict English Premier League match outcomes using first-half in-game statistics. The primary goal was to evaluate the predictive performance of different models, identify the most important features for prediction, and assess per-team accuracy. This was completed as the term project for a fundamentals of machine learning course.

Repository Contents

• DSCI_3415_Project.ipynb: The main Jupyter Notebook containing the complete analysis, model training, and evaluation.
• dsci_3415_project.py: The Python script version of the project code.
• DSCI 3415 Project Report.docx.pdf: The final project report detailing the methodology, results, and conclusions.
• DSCI 3415 Presentation.pptx: The slide deck for the project presentation.
• matches.zip: The dataset containing match information.
• teams.json: JSON file with team details.

Key Methods and Techniques

• Machine Learning Models: Implemented and compared three classification algorithms:
  • Logistic Regression
  • Random Forest
  • K-Nearest Neighbors (KNN)
• Model Evaluation: Performance was rigorously evaluated across 10 randomized trials to ensure robustness.
• Feature Importance Analysis: Investigated which first-half statistics (e.g., shots, possession, fouls) were most vital in predicting the final outcome.
• Per-Team Accuracy Visualization: Analyzed model performance on a per-team basis to identify strengths and weaknesses.

How to Reproduce This Work

1. Clone this repository to your local machine.
2. Ensure you have Python and Jupyter Notebook installed.
3. Install required libraries: pandas, scikit-learn, matplotlib, and seaborn.
4. Extract the matches.zip file to access the dataset.
5. Open and run the DSCI_3415_Project.ipynb notebook to reproduce the analysis.