Hierarchical Gaussian Process
A Bayesian nonparametric model that extends the standard Gaussian process with group-level structure, where each post-stratification cell has its own group effect and group effects are linked through a shared kernel structure. It is suited to survey weighting for structured, sparse populations and enables robust estimation of subgroup parameters.
Timeline (last 90 days)
This study proposes a survey-weighting framework based on hierarchical Gaussian process (HGP) regression, using kernel smoothing to share information across related units, and derives local equivalent weights.
The hierarchical Gaussian process introduces a group-level structure on top of the GP: each post-stratification cell has its own group effect, while the group effects are related to one another through a shared kernel structure.
Simulation experiments compared the performance of HGP and logistic regression: under nonparametric structures, HGP achieved lower population-weighted prediction error; while when the parametric model was correctly specified, logistic regression performed well.
HGP has advantages in sparse-data scenarios and can estimate subgroup parameters more stably.
The study applied the method to the "Future of Families and Child Wellbeing Study," demonstrating the operational workflow and effects of HGP in an actual survey.
With the growth of computational resources and the optimization of algorithms, nonparametric methods such as the hierarchical Gaussian process are expected to become one of the standard tools in survey statistics.
All Facts (6)
This study proposes a survey-weighting framework based on hierarchical Gaussian process (HGP) regression, using kernel smoothing to share information across related units, and derives local equivalent weights.
50%UnverifiedThe hierarchical Gaussian process introduces a group-level structure on top of the GP: each post-stratification cell has its own group effect, while the group effects are related to one another through a shared kernel structure.
50%UnverifiedSimulation experiments compared the performance of HGP and logistic regression: under nonparametric structures, HGP achieved lower population-weighted prediction error; while when the parametric model was correctly specified, logistic regression performed well.
50%UnverifiedHGP has advantages in sparse-data scenarios and can estimate subgroup parameters more stably.
50%UnverifiedThe study applied the method to the "Future of Families and Child Wellbeing Study," demonstrating the operational workflow and effects of HGP in an actual survey.
50%UnverifiedWith the growth of computational resources and the optimization of algorithms, nonparametric methods such as the hierarchical Gaussian process are expected to become one of the standard tools in survey statistics.
50%