Logistic Regression and Classification with non-Euclidean Covariates

02/23/2023
by   Yinan Lin, et al.
0

We introduce a logistic regression model for data pairs consisting of a binary response and a covariate residing in a non-Euclidean metric space without vector structures. Based on the proposed model we also develop a binary classifier for non-Euclidean objects. We propose a maximum likelihood estimator for the non-Euclidean regression coefficient in the model, and provide upper bounds on the estimation error under various metric entropy conditions that quantify complexity of the underlying metric space. Matching lower bounds are derived for the important metric spaces commonly seen in statistics, establishing optimality of the proposed estimator in such spaces. Similarly, an upper bound on the excess risk of the developed classifier is provided for general metric spaces. A finer upper bound and a matching lower bound, and thus optimality of the proposed classifier, are established for Riemannian manifolds. We investigate the numerical performance of the proposed estimator and classifier via simulation studies, and illustrate their practical merits via an application to task-related fMRI data.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
12/13/2017

Fractal dimension and lower bounds for geometric problems

We study the complexity of geometric problems on spaces of low fractal d...
research
11/22/2022

On Optimal Coreset Construction for Euclidean (k,z)-Clustering

Constructing small-sized coresets for various clustering problems in Euc...
research
05/13/2023

Tight and fast generalization error bound of graph embedding in metric space

Recent studies have experimentally shown that we can achieve in non-Eucl...
research
05/17/2022

Frank Wolfe Meets Metric Entropy

The Frank-Wolfe algorithm has seen a resurgence in popularity due to its...
research
08/05/2021

Evacuating from ell_p Unit Disks in the Wireless Model

The search-type problem of evacuating 2 robots in the wireless model fro...
research
05/31/2021

A Minimax Lower Bound for Low-Rank Matrix-Variate Logistic Regression

This paper considers the problem of matrix-variate logistic regression. ...
research
01/13/2021

Concurrent Object Regression

Modern-day problems in statistics often face the challenge of exploring ...

Please sign up or login with your details

Forgot password? Click here to reset