Research on Prediction of Landslide Displacement Based on BP Neural Network and D-S Evidence Theory
Kangdi Chen, Hongyang Liu, Xiaoling Tan · 2022 3rd International Conference on Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA) · 2022
Landslides are the most frequent disasters in nature and cause the greatest losses. Under the unpredictability and multi-factor influence of landslides, a suitable prediction model is the key to predicting landslides. In order to predict the displacement of the landslide, the displacement data of Zigui County were used to compare the displacement data of the monitoring points, and the ones with large deformation variables and many steps were selected for analysis. The cumulative displacement of landslide is divided into two displacements by a moving average method: trend displacement and periodic displacement. Prediction method: The cubic polynomial function predicts the displacement of the trend item, and the DS evidence theory integrates the prediction results of the BP neural network to predict the displacement of the periodic item. Finally, the predicted displacement of each sub-item is superimposed and compared with the actual accumulated amount. The results show that the predicted value of the BP-DS model is closer to the actual value, the error is smaller, and the model predicts the displacement with high feasibility and accuracy.