摘要
为探讨变权组合预测模型在沉降预测中的应用,以某大楼实测等时距沉降观测数据为研究对象,根据实测建筑物沉降数据所呈现的规律性,选取Logistic和Gompertz曲线模型对实测沉降数据进行了拟合;求取曲线模型的参数后,建立预测模型并进行预测。在此基础上,利用最小二乘法原理建立变权组合预测模型。根据各模型的预测结果及其计算出的模型精度显示,变权组合预测模型的拟合精度比Logistic和Gompertz曲线的拟合精度高,更适合于作为建筑物的沉降预测模型。
To investigate the application of the changeable weight combination model in building settlement predicting, the actually observed data of a building settlement has been taken as the research objective. According to the regularity shown in actually observed data, the Logistic and Gompertz curve models have been used to fit the observed settlement data, and after striking curve model parameters, the predicting model has been established and the prediction has been taken. On this basis, the least squares method principle has been used to establish the power mix change prediction model. The predicting accuracy of each model has been calculated according to the predicting results. The calculation shows that the fitting accuracy of the changeable weight combination predicting model is much more precise than that of the Logistic and Gompertz curve model fitting accuracy, which is more appropriate for the use of building settlement predicting.
出处
《测绘科学技术学报》
北大核心
2009年第2期118-120,124,共4页
Journal of Geomatics Science and Technology
