历史地震资料与地震的中长期预测

故障诊断模型的基本内容是根据动态系统的外部特征来判断系统内部是否发生故障及确定故障发生的部位、时间和大小。由于故障诊断技术在监测及诊断故障的思路上和地震预报有很多相似之处，因此，把故障诊断技术应用于地震预报是可行的。由于该模型与其他数学模型一样，需要有较多的学习过程，所以历史地震资料在该模型中有重要作用。而其中的鲁棒性故障诊断模型在抑制各子模型的个性，凸现其共性方面有其特有的性质，所以把它作为一种综合模型，能抑制各子模型的个性，突出在地震预测方面的共性，从而提高预测的精确性。本文根据一个实例，说明了这种综合性模型的可行性。在文章的最后，由信息量的分析，说明了模型的鲁棒性特征。

Historical Earthquake Data and Medium- and Long-Term Earthquake Prediction

The application of the Robust fault diagnosis model of automatic field in earthquake prediction has been shown. The basic idea of fault diagnosis model is to judge if the fault occur and to estimate it in space, time and level based on the external characteristics of the dynamical system. Because of the similarity in the idea of the detection and the diagnosis and between the fault diagnosis technique and the earthquake prediction, it should be feasibility in science that the fault diagnosis technique is applied in the earthquake prediction. Because the leaning process is necessary in this technique, which is as same as other mathematical model, the historical earthquake data would be taken as an important role in the model. The special feature of the robust fault diagnosis model, which is one of the fault diagnosis technique, is to give prominence to the generality of all sub-models in earthquake prediction and to restrain the individual feature of every sub-models. It would be possible to raise the reliability of the earthquake prediction.An example has explained the feasibility of the synthetic model. Finally,the robust feature of the model is given based on the analysis of the information value.