说明：  使用ARIMA-SVR组合预测的方式对数据进行预测，ARIMA与SVR分别对线性与非线性有较好的预测效果，因此通过SVR来预测ARIMA的残差使得预测结果更为精确
(Arima-svr combined forecasting method is used to forecast the data. ARIMA and SVR have better prediction effect on linear and nonlinear respectively. Therefore, the residual of ARIMA predicted by SVR makes the prediction result more accurate)