Statistical analysis of distance estimators with density differences and density ratios
Summary: Estimating a discrepancy between two probability distributions from samples is an important task in statistics and machine learning. There are mainly two classes of discrepancy measures: distance measures based on the density difference, such as the \(L_p\)-distances, and divergence measures based on the density ratio, such as the \(\Phi\)-divergences. The intersection of these two classes is the \(L_1\)-distance measure, and thus, it can be estimated either based on the density difference or the density ratio. In this paper, we first show that the Bregman scores, which are widely employed for the estimation of probability densities in statistical data analysis, allows us to estimate the density difference and the density ratio directly without separately estimating each probability distribution. We then theoretically elucidate the robustness of these estimators and present numerical experiments.
- Statistical properties of distance estimators
- Estimators of distance distributions for spatial patterns
- Estimating density ratio with application to discriminant analysis
- Density-difference estimation
- Statistical analysis of kernel-based least-squares density-ratio estimation
- On the distances between certain distributions in multivariate statistical analysis
- Relative Density-Ratio Estimation for Robust Distribution Comparison
- Estimation of distance functionals between density functions
- scientific article; zbMATH DE number 4001209
- f-Divergence Estimation and Two-Sample Homogeneity Test Under Semiparametric Density-Ratio Models
- Direct importance estimation for covariate shift adaptation
- Entropy and divergence associated with power function and the statistical application
- Estimating Divergence Functionals and the Likelihood Ratio by Convex Risk Minimization
- scientific article; zbMATH DE number 3322635 (Why is no real title available?)
- Improving predictive inference under covariate shift by weighting the log-likelihood function
- Inferences for case-control and semiparametric two-sample density ratio models
- Information Geometry of U-Boost and Bregman Divergence
- Proper local scoring rules
- Proper Scores for Probability Forecasters
- Robust and efficient estimation by minimising a density power divergence
- Robust parameter estimation with a small bias against heavy contamination
- Robust Truncated Hinge Loss Support Vector Machines
- Strictly Proper Scoring Rules, Prediction, and Estimation
- A Monte Carlo approach to quantifying model error in Bayesian parameter estimation
- Least informative distributions in maximum q-log-likelihood estimation
- Estimating number density N_V -- a comparison of an improved Saltykov estimator and the disector method.
- Estimating the Encounter Rate Variance in Distance Sampling
- A stand-alone analysis of quasidensity
- Relative Density-Ratio Estimation for Robust Distribution Comparison
- Density-difference estimation
- Outlier-robust parameter estimation for unnormalized statistical models
- Relative Entropy Gradient Sampler for Unnormalized Distribution
This page was built for publication: Statistical analysis of distance estimators with density differences and density ratios
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q296447)
