非等间隔加权G(1,1)模型在建筑物沉降预测中的应用

Application of non-equal interval weighted G(1,1) model in building settlement prediction

文中针对建筑物沉降监测数据量少、贫信息和非等时距等特点,引入单位时间差系数,将非等时距序列转化成等时距序列,建立非等间隔G(1,1)预测模型,依据监测数据对模型的贡献大小,引入权重矩阵以提高模型的预测精度。实践表明所建非等间隔加权灰色GM(1,1)模型的预测精度更高。

Aiming at the characteristics of less data,poor information and non-equal time interval of building settlement monitoring data,the unit time difference coefficient was introduced,and the non-equal time series were transformed into the equal time series to establish a non-equal interval G( 1,1) prediction model. The weight matrix was introduced to improve the prediction accuracy of the model according to the contribution size of the monitoring data for the model. Practice showed that the unequal interval weighted grey GM( 1,1) model had the higher prediction accuracy.