Smallest covering regions and highest density regions for discrete distributions
From MaRDI portal
Publication:2155016
Abstract: This paper examines the problem of computing a canonical smallest covering region for an arbitrary discrete probability distribution. This optimisation problem is similar to the classical 0-1 knapsack problem, but it involves optimisation over a set that may be countably infinite, raising a computational challenge that makes the problem non-trivial. To solve the problem we present theorems giving useful conditions for an optimising region and we develop an iterative one-at-a-time computational method to compute a canonical smallest covering region. We show how this can be programmed in pseudo-code and we examine the performance of our method. We compare this algorithm with other algorithms available in statistical computation packages to compute HDRs. We find that our method is the only one that accurately computes HDRs for arbitrary discrete distributions.
Recommendations
- Computing highest density regions for continuous univariate distributions with known probability functions
- Highest Posterior Density Credible Region and Minimum Area Confidence Region: The Bivariate Case
- On minquantile and maxcovering optimisation
- Maximizing cover probability by using multivariate normal distributions
- Optimal coverage of convex regions
Cites work
- scientific article; zbMATH DE number 3426675 (Why is no real title available?)
- scientific article; zbMATH DE number 44282 (Why is no real title available?)
- scientific article; zbMATH DE number 2107164 (Why is no real title available?)
- A genetic algorithm for the multidimensional knapsack problem
- A kernel estimator for discrete distributions
- Asymptotics and optimal bandwidth selection for highest density region estimation
- Bandwidth selection for kernel density estimators of multivariate level sets and highest density regions
- Convergence rates in nonparametric estimation of level sets
- Discrete associated kernels method and extensions
- Distribution-Free Prediction Sets
- Dynamic Programming and Strong Bounds for the 0-1 Knapsack Problem
- Estimation of a Convex Density Contour in Two Dimensions
- Improved interval estimation of long run response from a dynamic linear model: a highest density region approach
- Kernel methods for the estimation of discrete distributions
- Kernel smoothed probability mass functions for ordered datatypes
- Measuring mass concentrations and estimating density contour clusters -- An excess mass approach
- New trends in exact algorithms for the \(0-1\) knapsack problem
- Novel binary differential evolution algorithm for knapsack problems
- On finite sample properties of nonparametric discrete asymmetric kernel estimators
- On nonparametric estimation of density level sets
- Optimal rates for plug-in estimators of density level sets
- The highest confidence density region and its usage for joint inferences about constrained parameters
- Where are the hard knapsack problems?
Cited in
(2)
This page was built for publication: Smallest covering regions and highest density regions for discrete distributions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2155016)
