Download PDFOpen PDF in browserComparative Analysis of Machine Learning Algorithms for High-Accuracy Weather PredictionEasyChair Preprint 1593210 pages•Date: March 24, 2025AbstractAccurate weather prediction is crucial for applications like disaster preparedness, agriculture, aviation, and transportation planning. Traditional numerical weather models often struggle with long-term accuracy due to their sensitivity to perturbations and the complexity of atmospheric dynamics. Machine learning (ML) models have emerged as a promising alternative, leveraging historical weather data to enhance forecasting capabilities. This study evaluates multiple ML techniques using meteorological features such as temperature, precipitation, and wind speed. Among various ML techniques, ensemble based models demonstrated the most promise in capturing complex weather patterns, providing robust and reliable predictions. Additionally, Random Forest, Gradient Descent and Extreme Gradient Boosting (XGB) achieved superior performance in visibility prediction, surpassing traditional numerical weather models by reducing prediction errors. A comprehensive review of recent literature further emphasizes the rowing role of ML in weather forecasting, highlighting its advantages in handling large datasets and nonlinear relationships. Future work should focus on optimizing deep learning models, expanding training datasets, and incorporating additional meteorological factors such as humidity and atmospheric pressure to further enhance prediction accuracy. The study finds underscore the potential of MLbased approaches in advancing weather forecasting methodologies. By integrating these techniques with traditional meteorological models, more accurate and efficient weather prediction systems can be developed. Keyphrases: Meteorology Visibility Forecasting Atmospheric Data Forecasting Accuracy, Weather Prediction Machine Learning Ensemble Models, deep learning
|