【摘要】 对最小方差模型和泊松模型的原理进行阐述,并将模 型应用于对路基沉降进行预测。通过实例对两种模型的预测 结果进行对比分析。分析认为,采用泊松模型预测,用于建模 的实测数据越多,预测模型的精度将越高,误差可控制在10% 以内。采用最小方差模型预测,m0的选择比较重要,但缺少约 束,较泊松模型来说精度略显不足,应慎重对待。最后建议,为 提高预测精度,两种预测模型中新旧数据的权重应该有所不 同;应给两种预测模型赋以一定的权值,进行组合预测。更多还原

【Abstract】 Through describing the theory of Least Square Error Model and Poisson Model, this paper analyses the data from some railway subgrade settlement observation and predicts the settlement in the future by means of least square error model and Poisson model, and then gives the predicting results of two models. The result indicates that, using Poisson Model, the more observing data you have the better predicting precision you will obtain; using Least Square Error Model predict, how to choose m 0 is important. In the end, some valuable conclusions are drawn with contrasting the two forecasting models, and some useful suggestions are put forward.更多还原