Identification of slope stability based on the contrast of BP neural network and SVM

Abstract

Slope stability is always a very complex issue in engineering. Base on the theoretical analysis of BP neural network and support vector machine (SVM), some major factors which influence the slope stability are selected in soil slope and the slope samples are trained and identified. The identification rates of BP neural network and SVM both achieved 100%. In identification precision and elapsed time, the SVM can precisely identify the slope stability for the ability of outputting discrete values. By setting the threshold, BP neural network classifies the output results for identification. The difference between output values and the expected values is big. Also the results are unstable and the network training time is long. Selecting appropriate identification algorithms and determining an optimum one by comparative analysis is an effective identification method for slope stability.