This project analyzes feature importance in predicting oil price movements using financial market indicators. The goal is to understand which variables contribute most to predictive signals in a machine learning model.
The features include market and volatility indicators such as SPY, VIX, and lagged oil price values.
To evaluate the relative importance of different financial indicators in forecasting oil price movement using a machine learning-based approach.
- Feature engineering using lagged variables and moving averages
- Integration of market indicators (SPY, VIX, oil-related features)
- Machine learning model used for predictive analysis
- Feature importance extraction from trained model
- Performance evaluation on historical data
- Lagged oil features contribute significantly to predictive performance
- Volatility indicators (e.g., VIX) show meaningful influence on model output
- Equity market signals (e.g., SPY) provide additional but weaker predictive value
- Limited dataset range
- No incorporation of macroeconomic or geopolitical variables
- Model performance may vary across different market regimes
- Results are not intended as financial advice
paper.pdf→ Full research paper
This is an independent student research project exploring machine learning applications in financial time series analysis.