Topological classifier for detecting the emergence of epileptic seizures

TL;DR

This paper introduces a topological data analysis method using persistent entropy to classify EEG signals for epilepsy detection, demonstrating its effectiveness on Physionet data.

Contribution

It presents a novel application of topological data analysis with persistent entropy for classifying epileptic EEG signals.

Findings

Effective discrimination between healthy and epileptic EEGs.

Successful application on Physionet dataset.

Potential for improved seizure detection methods.

Abstract

In this work we study how to apply topological data analysis to create a method suitable to classify EEGs of patients affected by epilepsy. The topological space constructed from the collection of EEGs signals is analyzed by Persistent Entropy acting as a global topological feature for discriminating between healthy and epileptic signals. The Physionet data-set has been used for testing the classifier.