A Note on Smoothness Priors and Nonlinear Regression
本文推广了Shiller(1984)提出的非参数估计量，证明其与惩罚最小二乘估计和样条平滑直接相关，并简化了该估计量及其后验协方差的计算。
Abstract A generalization of a nonparametric estimator proposed by Shiller (1984) is considered. The estimator is shown to be directly related to penalized least squares estimation and spline smoothing. Some simplifications concerning the computation of Shiller's estimator and corresponding posterior covariances are indicated. Key Words: Bayesian predictionNonparametric regressionPartial splinesPenalized least squaresSpline smoothing