Indonesian Seismic Mitigation using Earthquake Predicted Artificial Intelligence Model

One way to reduce injuries during an earthquake is to determine the probability of an earthquake as early as possible so that rescue mitigation may be carried out. An artificial intelligence-based prediction model using a random forest regression algorithm was developed using a dataset of earthquake records in Indonesia from January 1900 to January 2022. The attributes used were date, location, latitude, longitude, hypocenter, magnitude, depth of the earthquake, maximum peak ground acceleration, modified Mercalli intensity, and site class. The label used to determine earthquake predictions was modified Mercalli intensity. The model was split into 80% training data and 20% testing data. The accuracy of the random forest regression model was 98.8%, and the error rate performance metrics mean absolute error and mean square error produce error values of 5.4% and 4.6%, respectively. This Artificial Intelligence-based earthquake prediction model is expected to provide input on the development of earthquake mitigation strategies in Indonesia. It is expected that there will be no more earthquake fatalities in the future.