Scholarlink

The inability to identify every seizure event and the complexity brought about by the high number of features utilized in classification algorithms are the main issues this research attempts to address. A unique method that greatly reduces the amount of characteristics and computing cost by using the Pythagorean mean as a feature for automatic seizure identification is put forth in order to get over these restrictions. In this paper, a KNN algorithm is presented for classifying epileptic EEG data into three levels of classes: two classes (normal and seizure), three classes (normal, interictal, and seizure), and five classes (normal with eyes closed, normal with eyes open, interictal from the opposite hemisphere, interictal from the epileptogenic zone, and seizure). Evaluation and comparison of the suggested algorithms’ performance with previously published techniques show how well they work to provide precise seizure detection while using the least amount of processing power.

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

Efficient Seizure Detection Using Pythagorean Mean and Simplified Feature Extraction for Multi-class EEG Classification

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Anand A. Khatri,
  • Rahul Patil,
  • Yogesh J. Pawar,
  • Vilas Ghonge,
  • Mahesh Bhandari,
  • Dattatray G. Takale,
  • Parikshit N. Mahalle,
  • Bipin Sule

摘要

The inability to identify every seizure event and the complexity brought about by the high number of features utilized in classification algorithms are the main issues this research attempts to address. A unique method that greatly reduces the amount of characteristics and computing cost by using the Pythagorean mean as a feature for automatic seizure identification is put forth in order to get over these restrictions. In this paper, a KNN algorithm is presented for classifying epileptic EEG data into three levels of classes: two classes (normal and seizure), three classes (normal, interictal, and seizure), and five classes (normal with eyes closed, normal with eyes open, interictal from the opposite hemisphere, interictal from the epileptogenic zone, and seizure). Evaluation and comparison of the suggested algorithms’ performance with previously published techniques show how well they work to provide precise seizure detection while using the least amount of processing power.