基于EMD的仪器稳定度组合预测

Combined Prediction Based on EMD for Instrument Stability

为更好地把握仪器的性能变化规律,规避其可能带来的各种风险,将预测理论应用到仪器稳定度预测中,建立了一种基于EMD-SVM的稳定度组合预测模型。首先利用EMD方法对稳定度数据进行分解,然后对分解得到的数据选择一种预测模型进行预测,最后再把所有这些分解数据的预测结果输入到SVM中进行组合预测。通过与移动平均模型、自回归积分滑动平均(ARIMA)模型和线性组合预测模型的预测结果相比较,验证了该方法的有效性。

In order to better grasp the regular pattern of the performance variation of the instruments, and to evade various risks may occur, the prediction theory is applied in the prediction of instrument stability, and the combined prediction model based on EMD-SVM is built. Firstly, the data of stability are decomposed by adopting EMD method; then a prediction model is selected from the data obtained by decomposition for predicting; finally all the predicted results of decomposed data are put into SVM for combined prediction. The comparison of the predictive results from moving average model, autoregressive integrated moving average （ ARIMA ） model and linear combination forecasting model, proves the effectiveness of this method.