Channel selection and creation algorithms for electroencephalography classification with HIVE-COTE

Rushbrooke, Aiden, Middlehurst, Matthew, Sami, Saber and Bagnall, Anthony (2025) Channel selection and creation algorithms for electroencephalography classification with HIVE-COTE. Corchado, Emilio, Quintián, Héctor, Troncoso Lora, Alicia, Pérez García, Hilde, Jove Pérez, Esteban, Calvo Rolle, José Luis, Martínez de Pisón, Francisco Javier, García Bringas, Pablo, Martínez Álvarez, Francisco, Herrero, Álvaro, Fosci, Paolo and Sérgio Filipe, Ramos (eds.) In Hybrid Artificial Intelligent Systems - 20th International Conference, HAIS 2025, Proceedings. vol. 16203 LNCS, Springer Cham. pp. 328-339 . (doi:10.1007/978-3-032-08462-0_26).

Abstract

Electroencephalography (EEG) is a crucial tool across neuroscience domains, including medical diagnostics, psychological research, and brain-computer-interfacing (BCI), due to its non-invasiveness, high temporal resolution, and cost-effectiveness. EEG data classification—the process of assigning predefined class labels to segments of EEG recordings based on patterns learned from training data—is challenging due to EEG’s high dimensionality, variability, and individual-specific differences. Recent research has shown that the time series classification algorithm HIVE-COTE v2.0 (HC2) is particularly effective at EEG classification, but that it is also orders of magnitude slower than algorithms from the EEG literature. We investigate ways of improving the run time of HC2 through channel selection and creation. We demonstrate that we can achieve accuracy that is not significantly different to full HC2 with up to 3 times faster runtime.

This record has no associated files available for download.

Contributors

Author: Matthew Middlehurst

Author: Anthony Bagnall

Download statistics