基于AR_SVR模型的时间序列预测算法的研究

掌握农产品未来价格变化趋势,有利于正确引导农业生产,提出一种基于自回归与支持向量回归(auto regressive and support vector regression,AR_SVR)模型的非平稳时间序列预测方法.首先,利用AR模型对非平稳时间序列进行季节差分和差分,使其具有平稳性,然后给平稳序列定阶,最后用SVR模型拟合平稳序列,回推得出原始序列的预测值.实验结果表明,AR_SVR模型预测值与真实值很接近,具有较好的预测效果.

Research of time-series forecasting algorithm based on AR_SVR model

To grasp the trend of agricultural prices is useful to proper guidance of agricultural production.Prediction method of non-stationary time series based on AR_SVR(auto regressive and support vector regression) model is proposed.Firstly,do seasonal difference and difference to non-stationary time series using AR model,which makes non-stationary time series stationary.Then the parameter of stationary series is determined.Finally,fit stationary series using SVR model and predictive value of original series is g...