Event Seismic Classification using Gaussian Processes

Seismic signals classification is important by itself in order to discover factual interactions between volcanic earthquakes and volcanic processes. In this paper, it is presented the application of Gaussian processes for seismic events classification registered at Nevado del Ruiz volcano. Feature extraction is accomplished using the coefficients of an autoregressive model, employed for the estimation of the power spectral density. The predictive distribution for classification is approximated using the Laplace method. Obtained performance is higher than the one obtained with an artificial neural network, the state of the art classifier for this kind of task.