Publications
A Deep auto-encoder based Framework for efficient weather forecasting
Dec 9, 2024Journal International Journal of Computational and Experimental Science and ENgineering (IJCESEN)
DOI 10.22399/ijcesen.429
Issue ISSN: 2149-9144
Volume 1053-1059
Weather forecasting has a plethora of benefits in different domains. Traditional weather forecasting approaches apply science and technology to predict weather conditions in a given place and time. With the emergence of artificial intelligence (AI), there are increased possibilities for weather forecasting research. Instead of ground-level observations, AI approaches learn from historical and current atmospheric data to develop predictions. We suggested a framework for autonomous weather forecasting based on deep learning. Our framework is a variant of the Convolutional Neural Network (CNN) model, which exploits the encoder and decoder to learn parameterizations from the given data and forecast weather. The proposed model can interpret spatial information associated with geopotential fields and automatically infers forecasting know-how with higher accuracy levels. A variable selection process is incorporated to determine geopotential height that impacts the weather conditions. We proposed an algorithm called Deep Weather Forecasting (DWF) to realize the proposed framework. Our empirical study has revealed that the proposed framework evaluates different deep learning models and compares their performance. Our deep learning models outperformed many existing regression models. U-Net showed the highest performance with the least MAE, 0.2268, compared to all other models.