Density-ratio matching under the Bregman divergence: a unified framework of density-ratio estimation

DOI10.1007/s10463-011-0343-8zbMath1440.62111OpenAlexW2155183960MaRDI QIDQ1926013

Takafumi Kanamori, Masashi Sugiyama, Taiji Suzuki

Publication date: 27 December 2012

Published in: Annals of the Institute of Statistical Mathematics (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1007/s10463-011-0343-8

zbMATH Keywords

logistic regressionBregman divergencedensity ratiokernel mean matchingKullback-Leibler importance estimation procedureleast-squares importance fitting

Mathematics Subject Classification ID

Density estimation (62G07) Asymptotic properties of nonparametric inference (62G20) Nonparametric estimation (62G05) Generalized linear models (logistic models) (62J12) Learning and adaptive systems in artificial intelligence (68T05) Statistical aspects of information-theoretic topics (62B10)

Related Items

Model-based policy gradients with parameter-based exploration by least-squares conditional density estimation, Positive-unlabeled classification under class-prior shift: a prior-invariant approach based on density ratio estimation, Estimating Density Ratio of Marginals to Joint: Applications to Causal Inference, Computational complexity of kernel-based density-ratio estimation: a condition number analysis, Density-Difference Estimation, Conditional Density Estimation with Dimensionality Reduction via Squared-Loss Conditional Entropy Minimization, Online Direct Density-Ratio Estimation Applied to Inlier-Based Outlier Detection, Direct Learning of Sparse Changes in Markov Networks by Density Ratio Estimation, Semi-supervised learning of class balance under class-prior change by distribution matching, Optimal adaptive estimation of the relative density, Change-point detection in time-series data by relative density-ratio estimation, Machine learning with squared-loss mutual information, Some Universal Insights on Divergences for Statistics, Machine Learning and Artificial Intelligence, Information-theoretic regularization for learning global features by sequential VAE, Unnamed Item, Information-Theoretic Representation Learning for Positive-Unlabeled Classification

Uses Software

• PDCO
• ElemStatLearn

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Direct importance estimation for covariate shift adaptation
• Statistical analysis of kernel-based least-squares density-ratio estimation
• Direct density-ratio estimation with dimensionality reduction via least-squares hetero-distributional subspace search
• Robust parameter estimation with a small bias against heavy contamination
• A compact formulation of an elastoplastic analysis problem
• Dual representation of \(\phi\)-divergences and applications.
• Improving predictive inference under covariate shift by weighting the log-likelihood function
• Semiparametric density estimation under a two-sample density ratio model
• Least angle regression. (With discussion)
• Least-squares two-sample test
• An introduction to variational methods for graphical models
• Test of homogeneity in semiparametric two-sample density ratio models
• A comparison of related density-based minimum divergence estimators
• 10.1162/153244302760185252
• Density Ratio Estimation in Machine Learning
• Least-Squares Independent Component Analysis
• Input-dependent estimation of generalization error under covariate shift
• Estimating Squared-Loss Mutual Information for Independent Component Analysis
• Atomic Decomposition by Basis Pursuit
• Robust and efficient estimation by minimising a density power divergence
• Inferences for case-control and semiparametric two-sample density ratio models
• Information Geometry of U-Boost and Bregman Divergence
• Estimating Divergence Functionals and the Likelihood Ratio by Convex Risk Minimization
• Elements of Information Theory
• Convex Analysis
• On Information and Sufficiency
• Robust Statistics
• The elements of statistical learning. Data mining, inference, and prediction
• Logistic regression, AdaBoost and Bregman distances