Support Vector Machines and Radon's Theorem

11/01/2020
by   Henry Adams, et al.
0

A support vector machine (SVM) is an algorithm which finds a hyperplane that optimally separates labeled data points in ℝ^n into positive and negative classes. The data points on the margin of this separating hyperplane are called support vectors. We study the possible configurations of support vectors for points in general position. In particular, we connect the possible configurations to Radon's theorem, which provides guarantees for when a set of points can be divided into two classes (positive and negative) whose convex hulls intersect. If the positive and negative support vectors in a generic SVM configuration are projected to the separating hyperplane, then these projected points will form a Radon configuration. Further, with a particular type of general position, we show there are at most n+1 support vectors. This can be used to test the level of machine precision needed in a support vector machine implementation. We also show the projections of the convex hulls of the support vectors intersect in a single Radon point, and under a small enough perturbation, the points labeled as support vectors remain labeled as support vectors. We furthermore consider computations studying the expected number of support vectors for randomly generated data.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
04/06/2018

Minimal Support Vector Machine

Support Vector Machine (SVM) is an efficient classification approach, wh...
research
05/13/2019

Exact high-dimensional asymptotics for support vector machine

Support vector machine (SVM) is one of the most widely used classificati...
research
06/24/2019

A Game-Theoretic Approach to Adversarial Linear Support Vector Classification

In this paper, we employ a game-theoretic model to analyze the interacti...
research
03/21/2019

Prescriptive Cluster-Dependent Support Vector Machines with an Application to Reducing Hospital Readmissions

We augment linear Support Vector Machine (SVM) classifiers by adding thr...
research
02/12/2016

General Vector Machine

The support vector machine (SVM) is an important class of learning machi...
research
05/28/2022

Temporal support vectors for spiking neuronal networks

When neural circuits learn to perform a task, it is often the case that ...
research
04/07/2019

Proposing a Localized Relevance Vector Machine for Pattern Classification

Relevance vector machine (RVM) can be seen as a probabilistic version of...

Please sign up or login with your details

Forgot password? Click here to reset