基于PCA-GA-BP神经网络的秦岭北麓周至-蓝田段山前地质灾害危险性预测

Risk Prediction of Geological Hazards in Front of Zhouzhi Lantian Section at the Northern Foot of Qinling Mountains Based on PCA-GA-BP Neural Network

地质灾害危险性预测工作对地质灾害防治、人民生命财产安全、地区开发建设等方面具有重要意义。由于常规预测模型存在一定不足,难以对地质灾害危险性做出准确预测。所以此类预测模型仍需进一步研究。建立了秦岭北麓周至-蓝田段山前地质灾害危险性PCA-GA-BP预测模型。将其训练10次得出最优预测结果和平均预测结果,分别与PCA-PSO-BP、RS-GA-BP预测模型进行对比。结果得出:文章提出的PCA-GA-BP的预测模型正确率较高,稳定性更好,为实际地质灾害危险性预测工程提供新思路。

Accurate prediction of geological hazard is of great significance to protect people's lives and properties,to carry out regional development and construction,and to prevent and control geological disasters.Due to certain shortcomings of conventional prediction models,it is difficult to make accurate predictions of geohazard hazards.Therefore,such prediction models still need further research.This paper establishes a PCA-GA-BP prediction model for the pre-hill geohazard hazard in the Zhouzhi-Lantian section of the northern Qinling Mountains,and compares it with PCA-PSO-BP and RS-GA-BP prediction models.The results conclude that the PCA-GA-BP prediction model proposed in this paper has higher accuracy and better stability,which provides new ideas for similar geohazard hazard prediction.