【摘要】 双曲线函数是拟合地基沉降曲线中常用的函数形式,其函数参数通常假定为确定性变量,采用最小二乘法进行估计。在整个待估计过程中,假定待估计参数为确定性变量。然而函数参数总是波动的,用随机变量分布代替其估计值更为合理,为此,将函数参数假定为随机变量,用Bayes理论进行估计,采用多层先验模型作为参数的先验分布,在似然函数中,沉降观测数据假定服从正态分布。在概率分布计算中,采用MCMC积分,所有计算均由软件WinBUGS完成。Bayes方法所得的估计是参数的概率分布,包含了比最小二乘法估计更为丰富的参数信息,因而比最小二乘法更具有优越性。更多还原

【Abstract】 The hyperbola function,which parameters are generally assumed as invariants estimated by least-square method,is a common function to simulate the foundation settlement curve.In this paper,the parameters are assumed as stochastic quantities,and evaluated by Bayesian interference.The hierarchical prior is taken as the parameters’ prior.The settlement distribution in likelihood function is supposed as normal distribution.After determined the posterior distribution of parameters,the predictive distribution of settlement can be calculated out.In the probability distribution computation,the MCMC integration is adopted,and all computations are conducted by freely software named WinBUGS.The results evaluated by Bayesian interference include more information about the parameters than that of least square method,so it is advantage of the least square method.更多还原