9 problems
- 0 votes0 replies1 view
Conjectured stability for unequal block sizes
Let the conditions of the stability theorem hold, but now for block sizes that are not necessarily equal. Define the mean block size and variance by … If the relative standar…
- 0 votes0 replies0 views
Existence of a non-degenerate optimal bandwidth for non-uniform spherical densities
Let be a density on the sphere, and let denote the sample size. A bandwidth is non-degenerate if it is strictly positive and minimizes the mean integrated squared error. No…
- 0 votes0 replies0 views
Li–Racine conjecture on bandwidth divergence for linear regressors
Li–Racine conjecture. The associated optimal bandwidth should diverge to infinity.
- 0 votes0 replies0 views
Adaptation of plug-in bandwidth selection methods to spatial trend regression
The setting is local polynomial regression for spatial data on , with bandwidths chosen for estimating the regression function and its derivatives from spatially depe…
- 0 votes0 replies0 views
Necessary moment condition for asymptotic optimality of CL under MDS errors
Let denote the cross-validation criterion used for bandwidth selection, let be the moment parameter in the assumptions, and let asymptotic optimality (a.o.) mean the…
- 0 votes0 replies0 views
Conjecture on the essential power adjustment for bandwidth selection consistency
Let denote the bandwidth parameter, let be the estimated residual variance associated with bandwidth , and consider the prior…
- 0 votes0 replies2 views
Conjecture on the asymptotic normality of the cross-validation bandwidth estimator
Bandwidth-normality conjecture. The normalized bandwidth error
- 0 votes0 replies0 views
Conjecture on honesty of rule-of-thumb confidence intervals
Honesty conjecture. For any , the resulting confidence interval is asymptotically honest over the intersection of and that appropriately defined set of regres…
- 0 votes0 replies0 views
Bandwidth-selection techniques for locally stationary nonparametric regression
The paper studies kernel estimators for a time-varying regression function with locally stationary covariates. In particular, its asymptotic theory gives bias and variance expressi…