LSTM and XGBoost Ensemble Model: An Approach for assessment and monitoring of Weather Prediction

Abstract

Managing the effects of climate change and extreme weather events that affect public safety, transportation, and agriculture etc requires accurate weather forecasting. This study offers a novel method for enhancing weather forecasting by utilizing Principal Component Analysis (PCA) for dimensionality reduction, Recursive Feature Elimination (RFE) for feature selection, Long Short-Term Memory (LSTM) networks for sequence modeling, and eXtreme Gradient Boosting (XGBoost) for predictive modeling. Comprehensive data collection, thorough preprocessing, feature selection, and the development of an ensemble model using LSTM and XGBoost are all part of the suggested methodology. This approach improves the ability to identify intricate, nonlinear relationships in meteorological data, leading to more accurate and reliable predictions. The model was evaluated using several metrics, achieving a high accuracy of 95.9% and an AUC of 0.83, demonstrating its effectiveness.