Predicting model and detecting factors causing rainfall using deep learning
Abstract
This study aims to employ deep learning algorithms to construct predictive models using real-world datasets containing indicators of rainfall. The objective is to determine the occurrence of rainfall at a specific point in time and to analyze the underlying factors contributing to its onset. Furthermore, the research is directed toward improving the accuracy of quantitative rainfall prediction for a given location and time. T h is study has developed a deep learning-based framework for weather forecasting with a particular focus on accurate rainfall prediction - a task that remains highly challenging not only for meteorological agencies in Vietnam but also for state-of-the-art forecasting systems worldwide. Using the collected dataset, we conducted descriptive statistical analyses to characterize its properties and investigated the parameters exhibiting correlations with rainfall events. Based on these findings, deep learning algorithms were applied to develop a classification model capable of predicting the probability of rainfall occurrence . The experimental results demonstrate that the proposed model can be applied to operational scenarios for forecasting rainfall at specific locations and times, utilizing rainfall indicators extracted from meteorological forecast databases. The outcomes of this research highlight the potential of artificial intelligence techniques in meteorological applications, offering the prospect of enhanced prediction accuracy and reduced risks associated with extreme weather phenomena.