Towards an Online Seizure Advisory System - An Adaptive Seizure Prediction Framework Using Active Learning Heuristics

Vignesh Raja Karuppiah Ramachandran (Corresponding Author), Huibert J. Alblas, Duc V. Le, Nirvana Meratnia

Research output: Contribution to journalArticleAcademicpeer-review

35 Downloads (Pure)

Abstract

In the last decade, seizure prediction systems have gained a lot of attention because of their enormous potential to largely improve the quality-of-life of the epileptic patients. The accuracy of the prediction algorithms to detect seizure in real-world applications is largely limited because the brain signals are inherently uncertain and affected by various factors, such as environment, age, drug intake, etc., in addition to the internal artefacts that occur during the process of recording the brain signals. To deal with such ambiguity, researchers transitionally use active learning, which selects the ambiguous data to be annotated by an expert and updates the classification model dynamically. However, selecting the particular data from a pool of large ambiguous datasets to be labelled by an expert is still a challenging problem. In this paper, we propose an active learning-based prediction framework that aims to improve the accuracy of the prediction with a minimum number of labelled data. The core technique of our framework is employing the Bernoulli-Gaussian Mixture model (BGMM) to determine the feature samples that have the most ambiguity to be annotated by an expert. By doing so, our approach facilitates expert intervention as well as increasing medical reliability. We evaluate seven different classifiers in terms of the classification time and memory required. An active learning framework built on top of the best performing classifier is evaluated in terms of required annotation effort to achieve a high level of prediction accuracy. The results show that our approach can achieve the same accuracy as a Support Vector Machine (SVM) classifier using only 20% of the labelled data and also improve the prediction accuracy even under the noisy condition.
Original languageEnglish
Article number1698
Number of pages30
JournalSensors (Switserland)
Volume18
Issue number6
DOIs
Publication statusPublished - 24 May 2018

Fingerprint

seizures
Problem-Based Learning
Adaptive systems
learning
Seizures
classifiers
predictions
Classifiers
Brain
ambiguity
Artifacts
brain
Quality of Life
Research Personnel
annotations
Pharmaceutical Preparations
Support vector machines
Heuristics
artifacts
drugs

Keywords

  • Health-care technology
  • EEG
  • Epilepsy
  • Signal processing
  • Machine Learning
  • Seizure Prediction
  • Implantable body sensor networks

Cite this

@article{73da96e8fce940a4b2241576a63efc3b,
title = "Towards an Online Seizure Advisory System - An Adaptive Seizure Prediction Framework Using Active Learning Heuristics",
abstract = "In the last decade, seizure prediction systems have gained a lot of attention because of their enormous potential to largely improve the quality-of-life of the epileptic patients. The accuracy of the prediction algorithms to detect seizure in real-world applications is largely limited because the brain signals are inherently uncertain and affected by various factors, such as environment, age, drug intake, etc., in addition to the internal artefacts that occur during the process of recording the brain signals. To deal with such ambiguity, researchers transitionally use active learning, which selects the ambiguous data to be annotated by an expert and updates the classification model dynamically. However, selecting the particular data from a pool of large ambiguous datasets to be labelled by an expert is still a challenging problem. In this paper, we propose an active learning-based prediction framework that aims to improve the accuracy of the prediction with a minimum number of labelled data. The core technique of our framework is employing the Bernoulli-Gaussian Mixture model (BGMM) to determine the feature samples that have the most ambiguity to be annotated by an expert. By doing so, our approach facilitates expert intervention as well as increasing medical reliability. We evaluate seven different classifiers in terms of the classification time and memory required. An active learning framework built on top of the best performing classifier is evaluated in terms of required annotation effort to achieve a high level of prediction accuracy. The results show that our approach can achieve the same accuracy as a Support Vector Machine (SVM) classifier using only 20{\%} of the labelled data and also improve the prediction accuracy even under the noisy condition.",
keywords = "Health-care technology, EEG, Epilepsy, Signal processing, Machine Learning, Seizure Prediction, Implantable body sensor networks",
author = "{Karuppiah Ramachandran}, {Vignesh Raja} and Alblas, {Huibert J.} and Le, {Duc V.} and Nirvana Meratnia",
year = "2018",
month = "5",
day = "24",
doi = "10.3390/s18061698",
language = "English",
volume = "18",
journal = "Sensors (Switserland)",
issn = "1424-8220",
publisher = "Multidisciplinary Digital Publishing Institute",
number = "6",

}

Towards an Online Seizure Advisory System - An Adaptive Seizure Prediction Framework Using Active Learning Heuristics. / Karuppiah Ramachandran, Vignesh Raja (Corresponding Author); Alblas, Huibert J.; Le, Duc V.; Meratnia, Nirvana .

In: Sensors (Switserland), Vol. 18, No. 6, 1698, 24.05.2018.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Towards an Online Seizure Advisory System - An Adaptive Seizure Prediction Framework Using Active Learning Heuristics

AU - Karuppiah Ramachandran, Vignesh Raja

AU - Alblas, Huibert J.

AU - Le, Duc V.

AU - Meratnia, Nirvana

PY - 2018/5/24

Y1 - 2018/5/24

N2 - In the last decade, seizure prediction systems have gained a lot of attention because of their enormous potential to largely improve the quality-of-life of the epileptic patients. The accuracy of the prediction algorithms to detect seizure in real-world applications is largely limited because the brain signals are inherently uncertain and affected by various factors, such as environment, age, drug intake, etc., in addition to the internal artefacts that occur during the process of recording the brain signals. To deal with such ambiguity, researchers transitionally use active learning, which selects the ambiguous data to be annotated by an expert and updates the classification model dynamically. However, selecting the particular data from a pool of large ambiguous datasets to be labelled by an expert is still a challenging problem. In this paper, we propose an active learning-based prediction framework that aims to improve the accuracy of the prediction with a minimum number of labelled data. The core technique of our framework is employing the Bernoulli-Gaussian Mixture model (BGMM) to determine the feature samples that have the most ambiguity to be annotated by an expert. By doing so, our approach facilitates expert intervention as well as increasing medical reliability. We evaluate seven different classifiers in terms of the classification time and memory required. An active learning framework built on top of the best performing classifier is evaluated in terms of required annotation effort to achieve a high level of prediction accuracy. The results show that our approach can achieve the same accuracy as a Support Vector Machine (SVM) classifier using only 20% of the labelled data and also improve the prediction accuracy even under the noisy condition.

AB - In the last decade, seizure prediction systems have gained a lot of attention because of their enormous potential to largely improve the quality-of-life of the epileptic patients. The accuracy of the prediction algorithms to detect seizure in real-world applications is largely limited because the brain signals are inherently uncertain and affected by various factors, such as environment, age, drug intake, etc., in addition to the internal artefacts that occur during the process of recording the brain signals. To deal with such ambiguity, researchers transitionally use active learning, which selects the ambiguous data to be annotated by an expert and updates the classification model dynamically. However, selecting the particular data from a pool of large ambiguous datasets to be labelled by an expert is still a challenging problem. In this paper, we propose an active learning-based prediction framework that aims to improve the accuracy of the prediction with a minimum number of labelled data. The core technique of our framework is employing the Bernoulli-Gaussian Mixture model (BGMM) to determine the feature samples that have the most ambiguity to be annotated by an expert. By doing so, our approach facilitates expert intervention as well as increasing medical reliability. We evaluate seven different classifiers in terms of the classification time and memory required. An active learning framework built on top of the best performing classifier is evaluated in terms of required annotation effort to achieve a high level of prediction accuracy. The results show that our approach can achieve the same accuracy as a Support Vector Machine (SVM) classifier using only 20% of the labelled data and also improve the prediction accuracy even under the noisy condition.

KW - Health-care technology

KW - EEG

KW - Epilepsy

KW - Signal processing

KW - Machine Learning

KW - Seizure Prediction

KW - Implantable body sensor networks

U2 - 10.3390/s18061698

DO - 10.3390/s18061698

M3 - Article

VL - 18

JO - Sensors (Switserland)

JF - Sensors (Switserland)

SN - 1424-8220

IS - 6

M1 - 1698

ER -