Scholarlink

<p>In this study, we introduce a robust and efficient diagnostic system for Parkinson’s disease (PD) based on an enhanced fuzzy k-nearest neighbor (FKNN) algorithm using firefly optimization (FFO). Our proposed system, termed FFO–FKNN, employs both continuous and binary versions of FFO for simultaneous parameter optimization and feature selection. Specifically, the continuous FFO adaptively determines the neighborhood size <i>k</i> and the fuzzy strength parameter <i>m</i> in the FKNN classifier, while the binary FFO selects the most discriminative feature subset for prediction. The FFO–FKNN model’s effectiveness was thoroughly assessed on five speech datasets, evaluating classification accuracy, recall, precision, F1-Score, and the area under the receiver operating characteristic (ROC) curve (AUC). Compared to previous methods, our system achieved the highest classification accuracy, with a mean accuracy of 100% through 10-fold cross-validation. This promising diagnostic system could potentially become a powerful tool for accurately diagnosing PD.</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

A New Evolutionary Fuzzy K-Nearest Neighbor with Application to Parkinson’s Diagnosis

  • 期刊论文

  • 发表日期：

    2026年2月6日

查看全文
  • Mohamed Samy,
  • Khaled Amin,
  • O. M. Abo-Seida,
  • Mina Ibrahim

摘要

In this study, we introduce a robust and efficient diagnostic system for Parkinson’s disease (PD) based on an enhanced fuzzy k-nearest neighbor (FKNN) algorithm using firefly optimization (FFO). Our proposed system, termed FFO–FKNN, employs both continuous and binary versions of FFO for simultaneous parameter optimization and feature selection. Specifically, the continuous FFO adaptively determines the neighborhood size k and the fuzzy strength parameter m in the FKNN classifier, while the binary FFO selects the most discriminative feature subset for prediction. The FFO–FKNN model’s effectiveness was thoroughly assessed on five speech datasets, evaluating classification accuracy, recall, precision, F1-Score, and the area under the receiver operating characteristic (ROC) curve (AUC). Compared to previous methods, our system achieved the highest classification accuracy, with a mean accuracy of 100% through 10-fold cross-validation. This promising diagnostic system could potentially become a powerful tool for accurately diagnosing PD.