Statistical Inference with Local Optima

DOI10.1080/01621459.2021.2023550arXiv1807.04431OpenAlexW4205549659MaRDI QIDQ6077585FDOQ6077585

Authors:

Publication date: 18 October 2023

Published in: Journal of the American Statistical Association (Search for Journal in Brave)

Abstract: We study the statistical properties of an estimator derived by applying a gradient ascent method with multiple initializations to a multi-modal likelihood function. We derive the population quantity that is the target of this estimator and study the properties of confidence intervals (CIs) constructed from asymptotic normality and the bootstrap approach. In particular, we analyze the coverage deficiency due to finite number of random initializations. We also investigate the CIs by inverting the likelihood ratio test, the score test, and the Wald test, and we show that the resulting CIs may be very different. We propose a two-sample test procedure even when the MLE is intractable. In addition, we analyze the performance of the EM algorithm under random initializations and derive the coverage of a CI with a finite number of initializations.

Full work available at URL: https://arxiv.org/abs/1807.04431

zbMATH Keywords

maximum likelihood estimationnonconvexEM Algorithmtwo-sample testgradient descent

Mathematics Subject Classification ID

Statistics (62-XX)

Cites Work

• Statistical analysis of finite mixture distributions
• Title not available (Why is that?)
• Asymptotic Statistics
• On the convergence properties of the EM algorithm
• Bootstrap methods: another look at the jackknife
• Title not available (Why is that?)
• Tests of Statistical Hypotheses Concerning Several Parameters When the Number of Observations is Large
• Mixture Densities, Maximum Likelihood and the EM Algorithm
• Title not available (Why is that?)
• All of Nonparametric Statistics
• On Estimation of a Probability Density Function and Mode
• On weak convergence and optimality of kernel density estimates of the mode
• Statistical guarantees for the EM algorithm: from population to sample-based analysis
• Morse Theory. (AM-51)
• Title not available (Why is that?)
• Note on the consistency of the maximum likelihood estimate for nonidentifiable distributions
• Title not available (Why is that?)
• A new regression model: modal linear regression
• Statistical inference using the Morse-Smale complex
• A Statistical Learning Approach to Modal Regression
• The landscape of empirical risk for nonconvex losses
• Robust topological inference: distance to a measure and kernel distance
• Gradient descent only converges to minimizers: non-isolated critical points and invariant regions

Cited In (2)

• Local statistical methods of knowledge formation
• Testing that a local optimum of the likelihood is globally optimum using reparameterized embeddings. Applications to wavefront sensing

This page was built for publication: Statistical Inference with Local Optima

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6077585)