Evolutionary Parameter Optimization on Neural Network Models for Earthquake Prediction

Earthquakes are a major obstacle to sustainable development, hindering social and economic growth. This study uses a model to predict the magnitude of earthquakes that occur from the Sunda Strait to Sumbawa Island. Earthquake prediction is important to take preventive measures and predict damage accurately. Several Earthquake Prediction (EQP) approaches have been proposed; however, these approaches only identify anomalies without distinguishing noise, thereby reducing the accuracy of predicting the probability of an earthquake occurring. The proposed model is a Neural Network (NN) optimized using evolutionary parameters to produce a lower and better error rate. Evolutionary Parameter Optimization was chosen because this parameter is superior in hyperparameter selection to achieve more optimal accuracy compared to other parameter models. Evolutionary Parameter Optimization was chosen because this parameter is superior in hyperparameter selection to achieve more optimal accuracy compared to other parameter models. This research aims to determine the best hyperparameter model to increase the accuracy of the Neural Network. The results of this study obtained the Root Mean Square Error (RMSE) value of the M 8 windowing combination using the Neural Network algorithm of 0.823. After increasing accuracy by optimizing using evolutionary parameters, the RMSE results obtained are 0.822. In this study, an increase in accuracy was obtained with a decrease in the RMSE value obtained by 0.001. Optimizing the Neural Network's evolutionary parameters improves the RMSE accuracy value so that the proposed model is better.