A Hybrid Early Earthquake Prediction Model Using IoT Networks and Deep Learning Algorithms

Abstract

Earthquakes are one of the most serious natural disasters that threaten human life and property, which makes their early prediction extremely important to reduce their devastating effects. This research aims to develop an intelligent earthquake prediction system based on the integration of Internet of Things (IoT) technologies with deep learning algorithms, specifically recurrent neural networks (RNN) and long-term short-term memory networks (LSTM). The methodology was developed by collecting Earthquake Data in real time using a network of IoT sensors, where historical real data from the year ( 1965 to 2021 ) was used to train the model and process this data using an optimized LSTM model with the use of activation functions (Relu,tanh , Sigmoid ) and optimization algorithms (Adam, RMSprop, SGD) . The model was developed using the ReLU activation function with the Adam optimization algorithm to improve the prediction accuracy with Learning Rate is equal to (0.001) and Epochs is equal to (200).. The experimental results showed a high efficiency of the proposed model, achieving a determination coefficient (R2).  By (0.9984) with a low MSE error rate of up to (0.0003), which confirms the system's ability to accurately predict earthquakes before they occur. This research makes an important contribution to the field of forecasting natural disasters using artificial intelligence and IoT technologies. These results confirm the possibility of relying on modern technologies in the development of earthquake early warning systems, which opens up new horizons for research in the field of Natural Disaster Risk Reduction and protection of vulnerable communities.