Accurate Precipitation Prediction using Deep Learning Neural Network Compared with Space Vector Machine

Aim: The proposed study evaluates Space Vector Machine (SVM) and Deep Learning Neural Network (DNN) rainfall prediction algorithms side by side. The algorithms' ability to reliably anticipate precipitation was used to assess their overall effectiveness. Methods and Materials We divide the study into two categories: group 1 consists of Space Vector Machines (SVMs) and group 2 consists of Deep Learning Neural Networks (DNNs). Pre-test G-power calculations with an alpha level of 0.05 yielded the necessary sample sizes for these cohorts. Each group receives 10 samples from the full datasets, and test-sample analyses are run to determine the statistical mean difference. The simulation results show that the new Space Vector Machine (SVM) obtains an accuracy of 84.24% in detecting rainfall, while the unique DNN method achieves an accuracy of 92.59% at a significance level of 0.0166
$(\mathbf{p\ 0.05})$
. By analyzing pictures from weather forecasts and comparing the results, the SVM and DNN algorithms were used to make predictions about past rainfall. When comparing the Deep Learning Neural Network (DNN) method to the tried-and-true Space Vector Machine (SVM) technique, the DNN approach emerges as the clear winner.