Study On GA-SVR Combined Model For Forecasting Landside Displacement Based On Hase-space Reconstruction

Slope as an important component of geotechnical environment has been a hot research object of scholars at home and abroad.In recent years,it is necessary to carry out slope deformation forecasting research.In this paper,support vector machine based on this emerging machine learning algorithm,around the slope deformation prediction problems to carry out research work.In this paper,firstly,the research status and significance of slope and SVM are expounded.The traditional Support Vector Machine(SVM)relies on the quality of the sample when dealing with short-period small sample data.The phase space of the sample based on chaotic dynamics Secondly,the genetic algorithm is used to solve the blindness problem of traditional support vector machine(SVM).Through genetic algorithm such as crossover,mutation and so on,the genetic algorithm is used to reconstruct the time series data by using GP algorithm and CC algorithm to determine embedding dimension and delay time.SVR combination model based on phase space reconstruction is obtained.Then Libsvm is compiled into MATLAB2102 by Microsoft Visual C ++,and then combined with extension toolbox to realize the support of traditional support vector machine and combined support vector machine Machine model of training,prediction.Finally,the combined model is applied to the prediction of deformation of a mine slope.It can be found by comparing the experimental results that the combined model has higher accuracy than the conventional SVR model,and the precision is improved by 30% to 50% The effect,the generalization ability all have the high superiority,has the certain project application and the promotion value.