This project analyzes soccer game data to predict match outcomes based on teams' historical performance in terms of expected goals (xG) and actual goals. It utilizes machine learning techniques, specifically K-Nearest Neighbors (KNN) and Rolling Forecasting, to provide predictions and insights.
A Machine Learning Approach for Predicting Outcomes of Premier League Soccer Matches Based on a Team’s Chance Creation and Quality of Finishing.
This research employs a machine learning approach to predict outcomes of Premier League soccer matches, focusing on the significance of a team’s ability to create scoring opportunities and their efficiency in finishing these chances. By analyzing data on chance creation and the quality of finishing, the study seeks to answer the question of which aspect is more crucial for a team’s success in the highly competitive environment of professional soccer. Utilizing metrics such as Expected Goals (xG) and actual goals scored, the paper investigates the relationship between the creation of high-quality scoring opportunities and their conversion into goals, offering insights into effective strategies for enhancing team performance.
- Check out our research paper here: Research Paper
- Explore the presentation here: Presentation Slides
Install the required Python libraries using pip:
pip install pandas matplotlib sklearn
#Run the Jupyter notebookLanden Fogle, University of Nebraska-Lincoln
Max Sievenpiper, University of Nebraska-Lincoln
Tage Zerby, University of Nebraska-Lincoln
Associated with RAIK 370H and Dr. Seth Polsley