Identifying anomalous signals in GPS data using HMMs: An increased likelihood of earthquakes?

摘要: A way of combining a hidden Markov model (HMM) and mutual information analysis is proposed to detect possible precursory signals for earthquakes from Global Positioning System (GPS) data. non-linear filter, which measures the short-term deformation rate ranges, introduced extract anomalous GPS measurements ground deformation. An HMM fitted filtered can classify data into different states form proxies elements earthquake cycle. Mutual then used examine whether any these possesses characteristics. The class identified by as having largest variation shows some hence considered ''precursory state''. performance forecasts assessed comparing decision rule (based on characteristics) with actual outcome.