Scholarlink

This work deals with EEG data analysis in studying meditation and predicting eye states by the use of a machine learning technique. In particular, the study focuses on the comparison of SVM and RVM classifiers. This paper compares the different features, channels, and spectrum indices using a single scaling condition. When evaluating the accuracy of the SVM and RVM classifiers in this study, performance was measured by changing window sizes and Sint. The accuracy of the SVM decreases as the window size becomes smaller. With larger window sizes, it resulted in higher accuracy. The highest accuracy of 95.12% resulted in a window duration of 2.5 min and a sample interval size of 16 s. However, the RVM classifier returned high accuracies for each window at all sampling intervals in contrast to the SVM classifier. However, a visible increase in accuracy from 93 to 94.23% was obtained when the time frame was brought down from 2 to 1 min. Both classifiers achieve results higher than 90%. SVM varies the range of parameter tuning and is shown to be sensitive at all times, while RVM has no sensitivity concerning change in sample rates. This paper allows for the determination of optimal scaling conditions in which, if known, it would improve classification outcomes in machine learning-based electroencephalogram studies. This would be possible if it were known which window sizes and sample interval values were more suitable.

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

Analysis of EEG Signals for Meditation and Prediction of Eye States Using Machine Learning Algorithms

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Parth Parmar,
  • Rituraj Jain,
  • P. Ramesh Babu,
  • Damodharan Palaniappan,
  • Mustafizul Haque

摘要

This work deals with EEG data analysis in studying meditation and predicting eye states by the use of a machine learning technique. In particular, the study focuses on the comparison of SVM and RVM classifiers. This paper compares the different features, channels, and spectrum indices using a single scaling condition. When evaluating the accuracy of the SVM and RVM classifiers in this study, performance was measured by changing window sizes and Sint. The accuracy of the SVM decreases as the window size becomes smaller. With larger window sizes, it resulted in higher accuracy. The highest accuracy of 95.12% resulted in a window duration of 2.5 min and a sample interval size of 16 s. However, the RVM classifier returned high accuracies for each window at all sampling intervals in contrast to the SVM classifier. However, a visible increase in accuracy from 93 to 94.23% was obtained when the time frame was brought down from 2 to 1 min. Both classifiers achieve results higher than 90%. SVM varies the range of parameter tuning and is shown to be sensitive at all times, while RVM has no sensitivity concerning change in sample rates. This paper allows for the determination of optimal scaling conditions in which, if known, it would improve classification outcomes in machine learning-based electroencephalogram studies. This would be possible if it were known which window sizes and sample interval values were more suitable.