季节指数及其在季节性监测资料分析和预报中的应用

基于灰色模型的诸多优点，作者选用GM(1，1)模型分析和预报形变监测序列。然而直接应用GM(1，1)灰色模型分析和预报具有季节性的监测序列时往往精度不高。因此，作者提出运用基于季节指数的“去季节波动”法与GM(1，1)混合建模，对监测资料进行分析与预报。基于均方差和平均绝对误差两个精度准则，作者对此方法与周期函数拟合模型进行了比较。结果表明，此方法提高了具有季节性波动监测序列的预报精度，且建模方法简便、快捷。

Seasonal Index and its Application in Analysis and Forecast of the Monitoring Series with Seasonality

Based on its characteristics, the grey forecasting model was applied to forecast the monitoring series with seasonality. However, the precision of GM(1,1) grey forecast model are not preferable to model the monitoring series with obvious seasonality. Therefore, the author proposes a hybrid model that combines the seasonal index-based deseasonalization method and GM(1,1) grey forecasting model to forecast monitoring series with seasonality characteristics. Based on two evaluation criteria, i.e., mean square error (MSE) and mean absolute error (MAE), we compare the performance of the hybrid model against other models, i.e., periodical function fitted model, with the monitoring series of a joint meter installed in a concrete structure. It proves that the hybrid model improve the forecasting precision. Besides, it diminishes modeling complexity.