回归和时间序列的组合模型在建筑物变形监测数据处理中的应用

摘要

为了提高变形预测的精度，获得可靠有效的预报模型，通过对回归模型残差项的进一步分析，利用Eview s对残差序列建立了时间序列模型，从而得到了回归和时间序列组合模型。利用该模型对一组实测数据进行分析、预测，并将变形数据和实测数据进行比较，证明了组合模型具有更好的预测精度和可靠性。%In order to improve the accuracy of predicting the deformation ,and abtain reliable and effective forecasting models ,We further analyzed the residuals of the regression model ,and built a time series model ,a combination of the regression and time series models with the rasidual se‐ries sing eviews .Using this model ,a set of measured data was analyzed and forecasted ,and the deformation data and measured data are compared .The results showed that the combination mod‐el had better prediction accuracy and reliability .