Estimation of MCS intensity for Italy from high quality accelerometric data, using GMICEs and Gaussian Naïve Bayes Classifiers

Abstract Macroseismic intensity provides a qualitative description of seismic damage. It can be associated with Ground Motion Parameters (GMPs), which are extracted in near real-time from instrumental recordings during an earthquake. Several formulations this empirical association exist literature for Italy, mainly focusing on the relationship between expressed Mercalli-Cancani-Sieberg (MCS) scale and peak ground acceleration or velocity. They usually form to Intensity Conversion Equations (GMICEs), treat as continuous quantity. We propose alternative approach, Gaussian Naïve Bayes (GNB) classifiers, allows correctly according its ordinal definition. As comparison, we also implement modified version standard GMICE approach. expand existing database GMP/MCS-intensity points new, high-quality accelerometric data recorded Italy period 2002 2016 resample by treating intermediate intensities half integer values (which not meaningful MCS description) both belong above below full classes assigned weight. result, estimate new set regression relations GNB probability distributions eight GMPs (peak acceleration, velocity, displacement, Arias Housner intensities, spectral at 0.3, 1.0 3.0 s). Results based PGA PGV most stable whole scale. models score better than GMICEs terms performance unseen classification scores.