Home / Current Issue / Paper 1712185
Neuroalert: AI-Based Seizure Prediction Software
Subject area: Science,Engineering and Technology · Area of research: Artificial Intelligence in Healthcare
DOI: https://doi.org/10.64388/IREV9I5-1712185
Abstract
Epileptic seizures occur unpredictably and often without prior warning, posing serious health risks to patients. Existing detection systems mainly rely on continuous EEG monitoring and biomedical sensors, which are expensive, require clinical environments, and are unsuitable for daily use. To overcome these limitations, this research introduces NEUROALERT, an AI-based seizure prediction software developed during its software phase without the use of physical sensors. The system simulates physiological parameters such as EEG, heart rate, skin conductance, and motion intensity, generating synthetic data for analysis. A Random Forest classifier is trained on this simulated dataset to recognize patterns indicating pre-seizure states. The model achieves an accuracy of 88%?92%, demonstrating the potential of software-based machine learning systems for reliable early seizure prediction. This work establishes a safe, cost-effective, and scalable foundation for developing future real-time wearable seizure alert devices.
Keywords
Artificial Intelligence (AI), Machine Learning (ML), Seizure Prediction, Biomedical Signal Simulation, Random Forest, Epilepsy Detection, Software-Based System.
How to cite this paper
@article{1712185,
author = {N. M. K. Ramalingam Sakthivelan, Ebin Andrews K, Dinesh M, Jagan N},
title = {Neuroalert: AI-Based Seizure Prediction Software},
journal = {Iconic Research And Engineering Journals},
year = {2025},
volume = {9},
number = {5},
pages = {1398-1411},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1712185.pdf},
abstract = {Epileptic seizures occur unpredictably and often without prior warning, posing serious health risks to patients. Existing detection systems mainly rely on continuous EEG monitoring and biomedical sensors, which are expensive, require clinical environments, and are unsuitable for daily use. To overcome these limitations, this research introduces NEUROALERT, an AI-based seizure prediction software developed during its software phase without the use of physical sensors. The system simulates physiological parameters such as EEG, heart rate, skin conductance, and motion intensity, generating synthetic data for analysis. A Random Forest classifier is trained on this simulated dataset to recognize patterns indicating pre-seizure states. The model achieves an accuracy of 88%?92%, demonstrating the potential of software-based machine learning systems for reliable early seizure prediction. This work establishes a safe, cost-effective, and scalable foundation for developing future real-time wearable seizure alert devices.},
keywords = {Artificial Intelligence (AI), Machine Learning (ML), Seizure Prediction, Biomedical Signal Simulation, Random Forest, Epilepsy Detection, Software-Based System.},
month = {November},
doi = {https://doi.org/10.64388/IREV9I5-1712185}
}