摘要
介绍了函数型数据半参数模型的估计问题,其中斜率函数满足单调性、凸凹性等形状约束条件.通过惩罚样条最小二乘估计推导出线性混合效应模型,进而提出了贝叶斯估计方法,并给出了马尔可夫链蒙特卡洛(MCMC)算法.模拟结果表明所提出的方法是有效的.
The paper introduces the estimation method for partial functional linear model when the slope function is subject to a variety of shape constraints such as monotonicity,convexity or concavity,etc.The linear mixed effects model is derived by the penalized splines least square estimate.Bayesian estimate is employed and Markov chain Monte Carlo(MCMC) algorithm is constructed.Simulation results show that the proposed method is effective.
作者
丁建华
DING Jian-hua(Department of Statistics,KLATASDS,Shanxi Datong University,Datong 037009,China)
出处
《数学的实践与认识》
2021年第18期178-184,共7页
Mathematics in Practice and Theory
基金
全国统计科学研究项目
教育部重点实验室开放课题。
