Harnessing Machine Learning for Reliable Weather Forecasting: Meteorological Impact on Sustainable Energy in Monterrey

In this study, we assess the performance of var-ious machine learning algorithms in weather forecasting for Monterrey, Mexico, a large metropolitan area with a mountain-influenced semi-arid climate whose geographical features significantly hinder weather prediction. To tackle this problem, we propose using heterogeneous models: Random Forest Regression, Support Vector Regression, Gradient Boosting Regression, and Long Short-Term Memory neural networks. Our research targets seven crucial weather parameters and utilizes lagged feature sets and an iterative prediction model to implement these ML models for forecasting weather conditions several days in advance. We compare the performance of our models via several metrics and statistical tests. Our findings reveal unexpected patterns, such as improved model accuracy over extended forecast periods. However, we find promising results for establishing the basis for future work on weather forecasting in mountain-influenced semi-arid climates.