利用相关因子预测中国大陆强震的支持向量机方法

Using correlation factors to forecast strong earthquakes in China′s mainland with Support Vector Machine

中国大陆地震活动与太阳活动、地球自转和全球地震活动之间存在较强的非线性关系。文中以太阳黑子数、地球自转速率变化数据和全球7级以上地震总应变释放量作为预测因子,使用支持向量机分类方法建模预测中国大陆年度地震强度,预测效果较好,表明支持向量机分类方法是一种较好的预测中国大陆地震活动强度的方法。

There is a strong and complicated nonlinear relation between earthquake activity in China＇s mainland and the Sun activity, the Earth rotation and global earthquake activity. In this paper, using the sunspot numbers, the Earth rotation rate variation data and the total strain release of the globe earthquakes of M≥7.0 as prediction factors, a classification model of Support Vector Machine is used to forecast earthquake intensity in China＇s mainland. The forecast effect of the model is relatively good, which shows the classification of Support Vector Machine is a relatively effective method for forecasting earthquake intensity in China＇s mainland.