测量仪器校准间隔的滚动灰色自助融合预测

针对测量仪器校准间隔的预测问题,在小样本条件下,采用滚动灰色自助融合模型进行校准间隔预测。滚动灰色自助融合模型综合灰色预测方法和自助再抽样方法,在灰微分方程建模时,通过自助再抽样,充分挖掘出系统信息。采用滚动灰色自助融合模型进行测量校准间隔预测,不仅能较准确地预测瞬时值,而且能够预测置信区间,克服了其他校准间隔预测模型仅仅预测瞬时值的缺点,降低了预测风险。实验表明,与其他校准间隔预测模型相比,滚动灰色自助融合模型的预测瞬时值、预测上限值和下限值都较好地描述出历史校准数据的波动趋势,预测可靠性更高,适合用于测量仪器校准间隔的预测。

Prediction of calibration interval for a measuring instrument using rolling grey bootstrap fusion model

A rolling grey bootstrap fusion model (RGBFM(1,1)) is proposed to predict calibration interval of a measuring instrument under small sample. The model combines GM(1,1) model with bootstrap method. Bootstrap re-sampling is used in the process of modeling the grey differential coefficient function to mine more information about systems. Both the instantaneous value and interval assessment values can be predicted using RGBFM(1,1), which can reduce prediction risk of calibration interval. In contrast, other prediction models only predict the instantaneous value. Experiments show that the RGBFM(1,1) can exactly describe the random wave of original sample data in prediction of instantaneous value, interval upper limit and lower limit, and has higher prediction reliability. Therefore, the RGBFM(1,1) is suitable for the prediction of calibration interval for a measuring instrument.