混沌和神经网络结合的滑坡预测方法

根据滑坡体运动的非线性动力学特性，应用混沌与神经网络相结合的预测方法，并采用数据实验的方法确定嵌入维数，建立滑坡预测的混沌模型。实验结果表明，用这种算法对清江茅坪滑坡体位移实际监测数据进行预测可达到较好的结果，能避免精确确定动力学方程和吸引子维数的困难，在滑坡体位移预测中具有实用性。

PREDICTION OF LANDSLIDE BASED ON CHAOS AND NEURAL NETWORKS

According to the nonlinear dynamical characteristics of landslide,with the prediction algorithms based on chaos and neural network,the embedding dimension is determined through data experiment,and the model of the chaotic process of landslide is set up. The experiment shows that this method can provide a satisfying result in the prediction of displacement monitoring data obtained in Maoping landslide of Qingjiang River. This method can avoid the difficulty to build the precise dynamics equations and determine the attractors dimension,and is of practicability in the prediction of landslide.