Ridge Fusion in Statistical Learning

This article proposes a penalized likelihood method to jointly estimate multiple precision matrices for use in quadratic discriminant analysis (QDA) and model-based clustering. We use a ridge penalty and a ridge fusion penalty to introduce shrinkage and promote similarity between precision matrix es...

Full description

Saved in:
Bibliographic Details
Published inJournal of computational and graphical statistics Vol. 24; no. 2; pp. 439 - 454
Main Authors Price, Bradley S., Geyer, Charles J., Rothman, Adam J.
Format Journal Article
LanguageEnglish
Published Alexandria Taylor & Francis 03.04.2015
American Statistical Association, Institute of Mathematical Statistics, and Interface Foundation of North America
Taylor & Francis Ltd
Subjects
Cluster analysis
Decision making models
Discriminant analysis
Estimating techniques
Joint inverse covariance matrix estimation
Matrix
Model-based clustering
Parameter estimation
Penalized Modeling for Regression and Classification
Semi-supervised learning
Studies
Online AccessGet full text

Cover

More Information
Summary:This article proposes a penalized likelihood method to jointly estimate multiple precision matrices for use in quadratic discriminant analysis (QDA) and model-based clustering. We use a ridge penalty and a ridge fusion penalty to introduce shrinkage and promote similarity between precision matrix estimates. We use blockwise coordinate descent for optimization, and validation likelihood is used for tuning parameter selection. Our method is applied in QDA and semi-supervised model-based clustering.
ISSN:1061-8600
1537-2715
DOI:10.1080/10618600.2014.920709