METHOD OF JUDGING EARTHQUAKE RISK TREND BY HUMAN-COMPUTER INTERACTION WITH A REAL EXAMPLE

Based on pattern recognition, using observation data for various precursors, including micro-seismicity, crustal deformation, earth resistivity, ground water level, radon content, we propose a method to judge earthquake risk trend by human-computer interaction. According to observation data before and after earthquakes and precursor phenomena predicated by models of earthquake forerunner, the quantitative index of precursor anomalies have been determined. Then the combination of the algorithm COEA-3 of pattern recognition with Fisher's linear discriminant function have been used to judge the coming earthquake risk trend. As an example, sample with earthquake, sample with no-earthquake and sample that needs prediction in North China have been discriminated from each other.