Improved Epilepsy Identification with XGBoost vs. Logistic Regression

Gadamsetty Hemanth, N. Navaprakash

2023

Abstract

This study aims to analyse and detect epilepsy with enhanced accuracy by implementing the XGBoost algorithm, comparing its performance against the Logistic Regression algorithm. Methodology: The research involves preprocessing and analysing a dataset containing instances of epileptic seizures. The dataset, sourced from CHB-MIT, is divided into two subsets, each consisting of 15 samples. The first subset applies the XGBoost algorithm, while the second employs the Logistic Regression algorithm for epilepsy disease detection. G power (80%) and alpha (0.005) values are determined for the study. Findings: The XGBoost algorithm achieves an impressive accuracy rate of 89%, surpassing the accuracy of the Logistic Regression algorithm at 74%. The observed difference is statistically significant, as confirmed by an independent sample t-test resulting in a p-value of 0.000 (p < 0.05). Conclusion: The study concludes that the innovative XGBoost algorithm excels with an accuracy of 89% compared to the Logistic Regression algorithm, establishing its effectiveness in epilepsy disease analysis and detection.

Paper Citation

in Harvard Style

Hemanth G. and Navaprakash N. (2023). Improved Epilepsy Identification with XGBoost vs. Logistic Regression. In Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics - Volume 1: AI4IoT; ISBN 978-989-758-661-3, SciTePress, pages 83-90. DOI: 10.5220/0012561800003739

in Bibtex Style

@conference{ai4iot23,
author={Gadamsetty Hemanth and N. Navaprakash},
title={Improved Epilepsy Identification with XGBoost vs. Logistic Regression},
booktitle={Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics - Volume 1: AI4IoT},
year={2023},
pages={83-90},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012561800003739},
isbn={978-989-758-661-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Artificial Intelligence for Internet of Things: Accelerating Innovation in Industry and Consumer Electronics - Volume 1: AI4IoT
TI - Improved Epilepsy Identification with XGBoost vs. Logistic Regression
SN - 978-989-758-661-3
AU - Hemanth G.
AU - Navaprakash N.
PY - 2023
SP - 83
EP - 90
DO - 10.5220/0012561800003739
PB - SciTePress