Access Application - Here
Access Notebook - Here
UFC Fight Predictor is a tool to predict the outcomes of upcoming or hypothetical UFC matchups.
- Data Acquisition: Leverages BeautifulSoup for efficient web scraping, gathering comprehensive fight data directly from relevant UFC sources.
- Data Processing: Employs a manual cleaning and processing approach within Jupyter Notebooks, ensuring data quality and relevance for model training.
- Machine Learning Models: Tests various predictive models, with Random Forest emerging as the top performer due to its 83% accuracy in predicting fight outcomes.
- User Interface: Features a user-friendly Streamlit application, allowing users to interactively select fighters and view predicted match outcomes. Results The Random Forest model achieved the highest accuracy, making it the backbone of our UFC Fight Predictor. This model's success lies in its ability to handle the complex nature of fight data and its proficiency in capturing the intricate relationships between various fighting attributes.
Interested in discussing this project further or collaborating on similar initiatives? Feel free to reach out and connect with me.