支持向量回归机在物价预测上的应用

Application of Support Vector Machine in Prediction of Consumer Price Index

把时间序列SVM预测模型应用于物价指数的预测,采用逐步递归的方法进行,同时注意尽量减少训练样本的浪费和充分挖掘SVM模型适合短期预测的潜力。分析结果表明,无论是拟合情况,还是预测值的检验和物价指数的实际规律来看,都有很高的精度,可以作为物价指数预测的一种行之有效的方法。

Using in prediction of Consumer Price Index, the recursive step-by-step method, taking care to minimize waste of training samples and fully tapping SVM model, the SVM time series forecasting model is suitable for short-term forecasts of potential. The results show that, whether it is fitting, or predictive value of the test and according to the actual price index of view, it has high accuracy.It can be an effective forecasting method as prediction of consumer price index.