The Zhongnan Road Slope Stability Analysis By Support Vector Machine

Influence factors of slope stability are complexity and uncertainty, and at the same timeslope failure is caused by a variety of elements of joint action. So it is necessary to find out thefactors of influence on slope stability and targete for governance. Traditional application in slopestability analysis including numerical simulation method, the limit equilibrium method andqualitative analysis method, have a large number of parameters or poor accuracy and othershortcomings. It is important to explore a less parameters and high accuracy method of slopestability analysis. The combination of fast messy genetic algorithm evolutionary support vectormachine algorithm, based on the existing data study, induction and analysis, could get theprediction of slope stability condition. The method based on slope stability of slope height,severe, slope angle, cohesion, friction angle and precipitation of six factors, considering thesmall amount of parameters can be achieved in cases with high accuracy.Support vector machine is a kind of structural risk minimization and statistical learningtheory based on machine learning methods. The main character is based on a small number ofdata samples and good generalization ability. Standard support vector machine algorithm alsohas a problem that the parameters have not a uniform standard and a fast calculation. This articaladopts fast messy genetic algorithm to be improved and develops a MATLAB calculationprogram. The calculation program is applied to Zhongnan road slope stability analysis, and theresults of the analysis and FLAC3D methods are compared. The program is feasibility in slopestability analysis of actual engineering.