Seismic data analysis in the Campi Flegrei caldera with Self-Organizing Map (SOM) and Multiscale Entropy (MSE)

This work proposes an innovative technique for seismic signal analysis in the Campi Flegrei caldera. The goal is to combine the Multiscale Entropy (MSE) algorithm and the Self-Organizing Map (SOM) network to address the challenge of detecting small earthquakes, which is challenging due to low signal-to-noise ratio and large, complex datasets. MSE measures data complexity across different time scales, while SOM enables unsupervised clustering for efficient interpretation of seismic data with reduced human bias. Our approach employs a four-month dataset from the V0102 station in the Pisciarelli area. By integrating these techniques, seismic data can be effectively clustered and interpreted, reducing human bias and enhancing the identification of key seismic events. The findings suggest that this methodology could prove beneficial for earthquake detection and classification within a given dataset.