灰色马尔科夫Verhulst动态模型在滑坡形变预测中的应用

为解决陈旧信息和波动性数据造成的传统灰色Verhulst模型预测精度较低的问题,提出一种利用滑动窗口和马尔科夫模型对原始灰色Verhulst模型改进的方法。通过长度可变的滑动窗口来实现数据的动态更新,使得灰色Verhulst动态模型的预测值更加接近最新的变化趋势。之后利用马尔科夫模型对得到的灰色Verhulst动态模型预测值进行修正,提高了模型的预测精度。实验结果表明,灰色马尔科夫Verhulst动态模型在滑坡形变预测中的预测平均相对误差相比于传统的灰色Verhulst模型降低了69. 6%,均方根差比降低了0. 39,小误差概率提高了0. 166 7。对于波动性较大的滑坡监测数据,灰色马尔科夫Verhulst动态模型预测精度优于传统灰色Verhulst模型。

Application of Grey Markov Verhulst Dynamic Model in Prediction of Landslide Deformation

In order to solve the problem that the traditional gray Verhulst model has low prediction accuracy caused by old information and volatility data, a method to improve the original gray Verhulst model by sliding window and Markov model is proposed. The dynamic update of the data is realized by a fixed-length sliding window, so that the predicted value of the gray Verhulst dynamic model is closer to the latest trend. Then, the Markov model is used to correct the predicted value of the gray Verhulst dynamic model, which improves the prediction accuracy of the model. The experimental results show that the predicted mean relative error of the grey Markov Verhulst dynamic model is reduced by 69.6% compared with the traditional gray Verhulst model in the landslide deformation prediction. The ratio of root mean square error of the improved model is 0.39 lower than that of traditional grey Verhulst model, and the small error probability is 0.1667 higher. For the volatility landslide monitoring data, the gray Markov Verhulst dynamic model prediction accuracy is better than the traditional gray Verhulst model.