Random Convolution Kernels with Multi-Scale Decomposition for Preterm EEG Inter-burst Detection

TL;DR

This paper introduces a multi-scale extension of the ROCKET method for preterm EEG inter-burst detection, combining high- and low-frequency features to improve classification accuracy efficiently.

Contribution

It proposes a novel multi-scale convolution kernel approach for EEG analysis, enhancing the existing ROCKET method with better performance in neonatal inter-burst detection.

Findings

Multi-scale ROCKET outperforms single-scale in MCC (0.859 vs. 0.841).

Method is fast to train and suitable for initial baseline in biomedical time-series.

Lags behind domain-knowledge-based methods but offers efficiency and simplicity.

Abstract

Linear classifiers with random convolution kernels are computationally efficient methods that need no design or domain knowledge. Unlike deep neural networks, there is no need to hand-craft a network architecture; the kernels are randomly generated and only the linear classifier needs training. A recently proposed method, RandOm Convolutional KErnel Transforms (ROCKETs), has shown high accuracy across a range of time-series data sets. Here we propose a multi-scale version of this method, using both high- and low-frequency components. We apply our methods to inter-burst detection in a cohort of preterm EEG recorded from 36 neonates <30 weeks gestational age. Two features from the convolution of 10,000 random kernels are combined using ridge regression. The proposed multi-scale ROCKET method out-performs the method without scale: median (interquartile range, IQR) Matthews correlation coefficient (MCC) of 0.859 (0.815 to 0.874) for multi-scale versus 0.841 (0.807 to 0.865) without scale, p<0.001. The proposed method lags behind an existing feature-based machine learning method developed with deep domain knowledge, but is fast to train and can quickly set an initial baseline threshold of performance for generic and biomedical time-series classification.