论文部分内容阅读
本文以模式识别为基础,利用地震活动、地壳形变、电阻率、地下水、水氡等几种前兆观测资料,提出了人机联合判断震情的方法。根据地震前后实际观测资料与孕震模式所预测的前兆变化,确定了单项前兆异常的定量标准,然后采用模式识别的 CORA-3算法与 Fisher 判别准则相结合的方法,由人机联合,最佳判断未来地震形势。作为方法的实例,本文对华北地区1969年以来的有震(M_s≥5.8)样本与无震样本进行学习,并对预测样本进行了地震危险性的判别。
Based on the pattern recognition, this paper presents a method of human-machine joint determination of earthquake conditions based on seismic precursors, crustal deformation, resistivity, groundwater and radon. Based on the observed data before and after the earthquake and the prediction of precursory changes in the seismogenic mode, the quantitative criteria for the single precursor anomalies were determined. Then the combination of the model-based CORA-3 algorithm and the Fisher criterion was used. Judge the future earthquake situation. As an example of the method, this paper studies the earthquakes (M_s≥5.8) samples and no-shock samples in North China since 1969, and distinguishes the seismic hazard of the predicted samples.