Scholarlink

Epilepsy is a chronic, incommunicable neuropsychiatric disorder marked by recurring epileptic seizures. Conventionally Electroencephalogram (EEG) is a test carried out by doctors for the diagnosis of epilepsy. It’s a prominent tool that measures electrical activity of brain by using electrodes attached to the patient’s scalp. These signals have complexity, noise, and non-linearity and produce voluminous data. The process of analyzing the EEG report manually is tedious and uncertain. This research article proposes an end-to-end solution to the problem using supervised machine learning techniques. The article is implemented on the Children’s Boston Hospital dataset that was rigorously studied on 23 subjects. The result for a ten-fold cross validation of the execution of the article is 82.72% for Random Forest and 62.76% for SVM. The experimental outcomes predict that the Random Forest shows better performance and has great potential in future epileptic detection studies.

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

Automated Epileptic Seizure Detection of EEG Signals Using Machine Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Shilpa Jain,
  • Vedant V. Gupta,
  • Ankur Singh Bist,
  • Mukesh Joshi,
  • Ashish Garg

摘要

Epilepsy is a chronic, incommunicable neuropsychiatric disorder marked by recurring epileptic seizures. Conventionally Electroencephalogram (EEG) is a test carried out by doctors for the diagnosis of epilepsy. It’s a prominent tool that measures electrical activity of brain by using electrodes attached to the patient’s scalp. These signals have complexity, noise, and non-linearity and produce voluminous data. The process of analyzing the EEG report manually is tedious and uncertain. This research article proposes an end-to-end solution to the problem using supervised machine learning techniques. The article is implemented on the Children’s Boston Hospital dataset that was rigorously studied on 23 subjects. The result for a ten-fold cross validation of the execution of the article is 82.72% for Random Forest and 62.76% for SVM. The experimental outcomes predict that the Random Forest shows better performance and has great potential in future epileptic detection studies.