Naive Bayes Classifier for Prediction of Volcanic Status in Indonesia

摘要

Disaster problems often occur in Indonesia is no exception disasters caused by volcanic eruptions, This is because Indonesia is the most active volcano in the world around 127 mountains. In determining the recommendation of volcanic status in Indonesia, the Center for Volcanology and Geological Hazard Mitigation (PVMBG) performs in two ways, namely visual observation, and seismic factors. This study will predict the status of volcanoes based on seismic factors. There are 5 criteria in the status recommendation are shallow volcanic earthquakes, deep tectonics, deep volcanic, earthquake blow and previous status. three status recommendations to be predicted are normal, alert and alert. In this study to predict the status of volcanoes used probabilistic reasoning, with the algorithm applied to predict the status of the volcano is the naive Bayes classifier. In addition, data validation was also done with k-fold cross-validation and then calculated the standard deviation. There is a 3-fold in this study. The results obtained an average accuracy of 79.91% and standard deviation obtained 3, 55%.