An Efficient Framework and Implementation of a Weather Prediction System

Abstract

The majority of today's weather forecasting studies have been focused on complex physical models. These models are usually run on hundreds of nodes in a High-Performance Computing system, which consumes a lot more power. Despite the employment of these costly and complex tools, projections are frequently incorrect due to inaccurate beginning conditions, measurements or a lack of understanding of atmospheric dynamics. Furthermore, solving complex models like this often takes a long time. The Internet of Things has helped any field that deals with technology. Using an IoT device, a prototype based on a machine learning approach is proposed in this study with an efficient framework, and implementation of an automated weather prediction system based on Artificial Neural Network algorithms was designed and developed. This system includes a technologically advanced irrigation system for our convenience. Using ANN in this research, the weather for the next day appears to have been predicted. The evaluation findings suggest that the model’s accuracy is sufficient for existing works and their approaches.