日本发生中强地震的灰色预测

GREY PREDICTION OF MODERATE AND STRONGEARTHQUMES IN JAPAN

本文应用灰色控制系统理论，选取１９８０年１月至１９８８年１２月日本的地震序列资料，将（６．０，７．０］级地震作为样本，建立了预报地震发震时刻的动态模型，并选取最佳模型对日本未来发生中强地震的时间进行了预测。另外，从函数变换的观点，对ＧＭ（１，１）模型进行了广义解释，指出对序列建模必须从满足光滑度的时刻计起。通过大量计算表明，用足够小量样本比大量样本建模，其拟合与外推精度要高。对此，从信息论的角度进行了剖析。

in this paper, the dynamic modcl of prodicting the origin time of carthquakes is established byusing the srey control system and choosing 6. 0<Ms<7. 0 seismic sequences from Jan. 1980 to Dec. 1988 inJapan as samples,and the best model is chosen to predict the occurrence of the moderate and strong carthquakcs inJapan. Morcovcr, in view of function transformation ,the model GM(1.1)is intrpreted in a broad sense and themodel cstablished for the scquencc mtist take time when the smooth degree is satisfied as the starting point.Morecalculatons indicate theat the fitting and extrapolated precisions of model cstablished“enough small number”ofsamplcs are better than that by “laorge number”of samplcs. This is anaysed from the view of information theory.