应变能释放随时间变化的聚类方法研究

CLUSTERING METHOD STUDIES ON VARIATION OF STRAIN ENERGY RELEASE WITH TIME

由于线性回归模型应用的广泛性，本文设定样品集的趋势是线性的。在用最优分割法对地震活动性分类研究的基础上引进线性模型，扩大了最优分割法在地震活动性分类中的应用，将聚类问题与建模问题有机地结合起来，去揭示地震活动性中的某种内在规律，这对于探索地震预报，研究地震发生的规律性有一定的实际意义。

It is assumed that the trends of samples are linear owing to the extensive application of linear regression model. A linear model is introduced to extend its applicasion based on the classification study of seismicity by optimum partition method. In order to study the inherent law of seismicity, clustering and model-making are combined together which has practical significance for exploring earthquake prediction and seismic pattern studies.