Saved in:
Bibliographic Details
Main Authors: Simpson, Andrew, Michael, Semhar
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.11404
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910911988498432
author Simpson, Andrew
Michael, Semhar
author_facet Simpson, Andrew
Michael, Semhar
contents In large-scale few-shot learning for classification problems, often there are a large number of classes and few high-dimensional observations per class. Previous model-based methods, such as Fisher's linear discriminant analysis (LDA), require the strong assumptions of a shared covariance matrix between all classes. Quadratic discriminant analysis will often lead to singular or unstable covariance matrix estimates. Both of these methods can lead to lower-than-desired classification performance. We introduce a novel, model-based clustering method that can relax the shared covariance assumptions of LDA by clustering sample covariance matrices, either singular or non-singular. In addition, we study the statistical properties of parameter estimates. This will lead to covariance matrix estimates which are pooled within each cluster of classes. We show, using simulated and real data, that our classification method tends to yield better discrimination compared to other methods.
format Preprint
id arxiv_https___arxiv_org_abs_2504_11404
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Statistical few-shot learning for large-scale classification via parameter pooling
Simpson, Andrew
Michael, Semhar
Methodology
In large-scale few-shot learning for classification problems, often there are a large number of classes and few high-dimensional observations per class. Previous model-based methods, such as Fisher's linear discriminant analysis (LDA), require the strong assumptions of a shared covariance matrix between all classes. Quadratic discriminant analysis will often lead to singular or unstable covariance matrix estimates. Both of these methods can lead to lower-than-desired classification performance. We introduce a novel, model-based clustering method that can relax the shared covariance assumptions of LDA by clustering sample covariance matrices, either singular or non-singular. In addition, we study the statistical properties of parameter estimates. This will lead to covariance matrix estimates which are pooled within each cluster of classes. We show, using simulated and real data, that our classification method tends to yield better discrimination compared to other methods.
title Statistical few-shot learning for large-scale classification via parameter pooling
topic Methodology
url https://arxiv.org/abs/2504.11404