Rainfall Forecasting of Landslides Using Support Vector Regression

Natural disasters caused by landslides cause many losses in agricultural production, life, and property. The Sustainable Development Research Center of Slope Land (SDRCSL) built a weather station with inclinometers to provide early warning for landslides in Huafan University. The collected data from the weather station include insolation, rainfall, temperature, humidity, the speed of wind, and the tilt of the slope. In this paper, rainfall forecasting of landslides using support vector regression (SVR) is proposed. In the proposed method, SVR forecasts rainfall of landslides on the Apache Spark platform. Apache Spark can carry out parallel in-memory large-scale data processing. Compared with other methods, the results provide the smallest root mean square error (RMSE) for rainfall forecasting of landslides.