Earthquake is Predicted

Abstract

Abstract Earthquake prediction is a global challenge. Seismologists have established a large number of observation stations in active seismic areas that provide massive, continuous, and complete geomagnetic and geoacoustic data from different regions. Based on these data, we have developed a convolutional neural network earthquake forecasting model to achieve short-term earthquake prediction. After normalizing geomagnetic and geoacoustic observation data, we randomly divided the data into training and testing groups, inputted the training group into the convolutional neural network model for training, and used the resulting model to test testing group and calculate the accuracy. Our research shows that the model has approximately 81% accuracy(17.7% higher than 620 groups of researchers using same datasets). The model is suitable for integrating geomagnetic and geoacoustic data and has great potential for improving the accuracy of earthquake prediction in China, and all other regions if datasets are available.