Imputation estimators for unnormalized models with missing data
Imputation estimators for unnormalized models with missing data
Date
2019-03-12
Authors
Uehara, Masatoshi
Kim, Jae Kwang
Matsuda, Takeru
Kim, Jae Kwang
Kim, Jae Kwang
Matsuda, Takeru
Kim, Jae Kwang
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Kim, Jae Kwang
Person
Research Projects
Organizational Units
Statistics
Organizational Unit
Journal Issue
Series
Department
Statistics
Abstract
We propose estimation methods for unnormalized models with missing data. The key concept is to combine a modern imputation technique with estimators for unnormalized models including noise contrastive estimation and score matching. Further, we derive asymptotic distributions of the proposed estimators and construct the confidence intervals. The application to truncated Gaussian graphical models with missing data shows the validity of the proposed methods.
Comments
This pre-print is made available through arxiv: https://arxiv.org/abs/1903.03630.