A neurocomputing approach for anomaly detection of Mt. Merapi monitoring activity

Abstract

Monitoring volcanic activity is extremely important to detect anomalies that may changes in the activity of Mt. Merapi. In this paper we proposed Multi Layer Perceptron (MLP) method to detect anomaly and to determine activities of each quake in seismic data. This method that has been developed in this research has been tasted against such several types of quakes as volcanic A (VA), volcanic B (VB), multiphase (MP), and avalance using data of the same time period. The experimetal results showed an average accuracy of 81,7 % in determining the activity of each quake type of Mt. Merapi seismic activity.