Scholarlink

<p>This research presents an innovative approach designed for the meditation domain, specifically addressing classification involving expert meditators, novice meditators, and a control (lacks meditation experience). Electroencephalography (EEG) has emerged as a promising tool in the realm of meditation research due to its accessibility and simplicity of implementation. The research purpose is twofold: firstly, to develop a robust feature selection methodology to capture pertinent patterns within meditation data, and secondly, to optimize Support Vector Machine (SVM) parameters, including gamma, C, and kernel type, for enhanced classification accuracy. The raw EEG data is rigorously pre-processed and then subjected to feature extraction phase. This phase extracts non-linear and statistical features tailored to meditation data analysis. The Harry Hawks Optimization algorithm is employed for feature selection, strategically identifying the most discriminating features. Additionally, parameter optimization for the SVM classifier is achieved through grid search, facilitating the identification of optimal hyper-parameters to mitigate over-fitting and improve model generalization. Evaluation entails a five-fold cross-validation process, ensuring robustness and reliability in model assessment. Results with 98.71% and 96.02% accuracy values with two datasets, demonstrate the efficacy of the proposed approach in constructing accurate classification models related to meditation practitioner categories.</p>

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

Enhancing Classification in the Meditation Domain: Harry Hawks Optimization Approach for Feature Selection and SVM Parameter Optimization with Grid Search

  • 期刊论文

  • 发表日期：

    2025年6月16日

查看全文
  • Soniya Shakil Usgaonkar,
  • Damodar Reddy Edla,
  • R Ravinder Reddy

摘要

This research presents an innovative approach designed for the meditation domain, specifically addressing classification involving expert meditators, novice meditators, and a control (lacks meditation experience). Electroencephalography (EEG) has emerged as a promising tool in the realm of meditation research due to its accessibility and simplicity of implementation. The research purpose is twofold: firstly, to develop a robust feature selection methodology to capture pertinent patterns within meditation data, and secondly, to optimize Support Vector Machine (SVM) parameters, including gamma, C, and kernel type, for enhanced classification accuracy. The raw EEG data is rigorously pre-processed and then subjected to feature extraction phase. This phase extracts non-linear and statistical features tailored to meditation data analysis. The Harry Hawks Optimization algorithm is employed for feature selection, strategically identifying the most discriminating features. Additionally, parameter optimization for the SVM classifier is achieved through grid search, facilitating the identification of optimal hyper-parameters to mitigate over-fitting and improve model generalization. Evaluation entails a five-fold cross-validation process, ensuring robustness and reliability in model assessment. Results with 98.71% and 96.02% accuracy values with two datasets, demonstrate the efficacy of the proposed approach in constructing accurate classification models related to meditation practitioner categories.