Gas Price Predictor

Built for an AI class at the University of Houston, as part of a 4-person team comparing machine learning approaches to price forecasting. My contribution: one of the project’s LSTM models, plus data collection and research.

• Compared LSTM, GRU, and Random Forest models on historical USOIL prices
• Technical indicators as features: moving averages, RSI, Bollinger Bands, MACD
• Separate Random Forest model for weekly Houston gas price changes
• GRU outperformed LSTM on this dataset (lower error)
• Key finding: both models struggled with sudden price spikes driven by real-world events