AHP与GA-SVM耦合模型在滑坡预警中的应用*

为了减少滑坡造成的损失，提高滑坡预测的准确性，通过搭建灾害模拟平台获得滑坡的实验数据，在获得多组模拟实验数据后，分析各变量的特性。首先，通过层次分析(Analytic Hierarchy Process，AHP)算法，对滑坡进行危险度划分；然后，通过支持向量机(Support Vector Machine，SVM)建立模型，遗传算法(Genetic Algorithm,GA)再优化SVM参数，提出1种层次分析法与GA-SVM相耦合的模型。研究结果表明:AHP方法划分后的数据，通过GA与SVM结合建立的模型精度较好，实验预测结果与实际结果较为吻合，与单一SVM相比，精度更高，结果更好，更加适用于多变量的复杂非线性滑坡预警。

Application of AHP and GA－SVM coupling model in landslide warning

In order to reduce the loss caused by landslides and improve the accuracy of landslide prediction,the experimental data of landslide were obtained through establishing the disaster simulation platform,and the characteristics of each variable were analyzed after obtaining multiple sets of simulation experimental data.Firstly,the risk classification of landslide was conducted by using the analytic hierarchy process (AHP) algorithm.Then a model was established through the support vector machine (SVM),and the SVM parameters were re optimized through the genetic algorithm (GA).Finally,a coupling model of AHP and GA-SVM was proposed.The results showed that for the data after the classification by AHP method,the model with the combination of GA and SVM had better accuracy,and the experimental prediction results were in good agreement with the actual results.Compared with single SVM,it has higher accuracy and better results,and is more suitable for multivariable complex nonlinear landslide warning.