Ensemble Temperature Forecasting
Published:
Keywords: Time-Series Analysis, Ensemble Learning, Data Engineering
Project Details
Designed a robust forecasting system for Montreal’s hourly temperature fluctuations, addressing the challenge of rapid local weather shifts.
- Methodology: Implemented a multi-model framework comparing single-model baselines (Linear Regression, LSTM) against Ensemble Methods.
- Engineering: Automated the data processing pipeline to clean and normalize raw meteorological data for real-time evaluation.
- Results: The ensemble approach demonstrated superior stability and reduced Mean Squared Error (MSE) compared to individual regressors in high-variance weather conditions.