Scholarlink

A stroke has the potential fatal if neglected among teenagers and older people are more likely to suffer from stroke disability that could result in a variety of societal and financial challenges. It is shown that individuals who have a stroke display aberrant biological signals. A stroke demands quick action which takes more time to identify and address a condition that might hurt an individual's prognosis. To overcome this issue, we have proposed the Ensemble Bee Colony intelligence multi-featured adaptive convolutional neural network (EBCI-MACNN) to predict stroke illness. In this paper, we have used a wide range of health-related information including illnesses and environmental characteristics. We collected the electroencephalogram (EEG) dataset and preprocessed using min-max normalization and then feature extracted through a discrete wavelet transform (DWT). Stroke can be affected and deadly for older people as they have less immune power in their bodies. These novel strategies were able to provide quick and more reliable stroke alerts that would enable preventive medical measures. The proposed method obtained 96% of accuracy, 98% of precision, 95% of F1 score, and 96% of specificity.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Revolutionizing Stroke Prediction Framework: A Deep Learning Approach Using Real-Time EEG Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Venkat Namdev Ghodke,
  • A. K. M. B. Hossain,
  • Rakesh Thoppaen Suresh Babu,
  • R. Mohan Kumar

摘要

A stroke has the potential fatal if neglected among teenagers and older people are more likely to suffer from stroke disability that could result in a variety of societal and financial challenges. It is shown that individuals who have a stroke display aberrant biological signals. A stroke demands quick action which takes more time to identify and address a condition that might hurt an individual's prognosis. To overcome this issue, we have proposed the Ensemble Bee Colony intelligence multi-featured adaptive convolutional neural network (EBCI-MACNN) to predict stroke illness. In this paper, we have used a wide range of health-related information including illnesses and environmental characteristics. We collected the electroencephalogram (EEG) dataset and preprocessed using min-max normalization and then feature extracted through a discrete wavelet transform (DWT). Stroke can be affected and deadly for older people as they have less immune power in their bodies. These novel strategies were able to provide quick and more reliable stroke alerts that would enable preventive medical measures. The proposed method obtained 96% of accuracy, 98% of precision, 95% of F1 score, and 96% of specificity.