Time Series Prediction of Landslide Displacement Using SVM Model: Application to Baishuihe Landslide in Three Gorges Reservoir Area, China

Time series analysis has the ability to forecast the evolve trend of complex systems, which has been the issue in the research of landslide displacement dynamic forecasting. The Support Vector Machine (SVM) regression, we proposed, has been applied in Baishuihe landslide in Three Gorges Reservoir Area, China. The Oracle Data Mining (ODM) PL / SQL API have been introduced to build the SVM regression model for data mining process. The data was divided into two parts, wherein the first 36 months data used for training, and the other 6 months data used for validation. The results show that the error rate of the previous 5 was controlled within 8% and the accuracy of the 6th is 84.1%, which indicates SVM regression is reliable to calculate the displacement factors and can be used in short term prediction of landslide monitoring data.