四川三个地震带逆向时间序列的算法复杂性与地震预报研究

Research on The Relationship Between The Algorism Complcxity of Reversed Time Sequence And The Prediction Of Earthquake Occurred Along The Three Seismic Belts In Sichuan

本文利用四川三个重要地震带：鲜水河地震带，松潘、龙门山地震带和名山－马边－昭通地震带二十多年地震资料，以一年为时间窗，半年为滑动步长，按２．５≤Ｍ≤３．５，３．５≤Ｍ≤４．５和Ｍ≥４．５三个震级档研究了应用逆向时间序列的算法复杂性预测三带未来半年发生５级以上地震的可能性，得到了与正向时间序列算法复杂性大体一致或稍高的预报效能。它们的预报评分Ｒ值都明显高于置信度９７．５％的区分随机盲报的临界评分值Ｒ＿０。

Using the seismic data more than 20 years of the main three seismic belts,which are the Xi-anshuihe seismic beklt、the Songpan-Longmenshan seismic belt and the Mingshan-Mabian- Zhao-tong seismic belt,we took a year as a time window and a half year as a moving step according tothe three grades of magnitude 2.5≤M<3.5,3.5≤M<4.5 and M≥4.5.We researched theprobability of predicting earthquake which will occur along the three belts in future and whosemagnitude is more than 5. 0 with the algorism complexity of reversed time sequence in this paper. We have obtained the prediction effect which is generally consistent with the result using the al- gorism complexity of positive time sequence or is higher than,Their critical markRof predictionis obviously higher than the critical mark R_0 of stochastic blind prediction. Confidence level is97.5%。