Civic Centered Heuristic Early Warning System Fashioned using Artificial Intelligence and Internet of Things

Nowadays, countries are suffering from a significant problem of natural disasters worldwide. Natural disasters happen due to environmental imbalance. The previously developed models do not have substantial results. The early warning system (EWS) comprises artificial intelligence (AI) and the internet of things (IoT) technologies. The EWS is tracing the climatic/weather conditions and accordingly floats an intimation to humans geographically through network-connected mobile devices. EWS gets trained using a 2D convolution neural network (CNN) of deep learning (DL) algorithms on collected data from different weather sensors. It classifies the weather conditions accurately. The proposed design is an EWS prediction model for detecting future natural disasters by following past and current climatic data details. It attains 93% training accuracy, 90% validation accuracy, 22% training losses, and 34% validation losses approximately. Also, to measure the model’s performance for multiclassification on the validation dataset, find the precision, recall, and f1-score for each class, respectively. Then, calculate the accuracy, macro average (macro avg) and weighted average (weighted avg) on the whole testing dataset. All of the following results are explained in the classification report section. Simultaneously, this system provides a warning message to society geographically through IoT devices.