Scholarlink

<p>The utilization of Data Mining (DM) and Machine Learning (ML) approaches in Alzheimer Disease (AD) diagnosis has lately gained a lot of attention. However, utmost of these studies still required improvement since either they were evaluated utilizing insufficient assessment-measurements; or they were not statistically-evaluated; or both. Recently, the PSO-ELM (Particle Swarm Optimization-Extreme Learning Machine) is considered one-of-the-most well-known and effective ML methods, it has seen as a reputable and efficient technique for categorizing-data, however it hasn’t been employed in AD diagnosis issue. Thus, this study proposes the PSO-ELM model in-order-to lift the accuracy rate of AD diagnosis. The PSO-ELM model has the capability to (i) ability to be utilized on both (binary/multi-class) classification, (ii) prevent overfitting; and also (iii) it has a similar capability to kernel-based SVM (Support Vector Machine) and functions with a structure of neural-network. In this paper, the DARWIN (Diagnosis AlzheimeR WIth haNdwriting) dataset was utilized to assess the PSO-ELM model performance. The experimentations results have demonstrated the distinct performance of the proposed PSO-ELM model, which reached an average of F-Measure 98.58%, recall 100.00%, specificity 97.60%, precision 97.59%, G-Mean 98.60%, MCC 97.22%, and accuracy 98.57%. This designates that the PSO-ELM model is a reliable method for AD diagnosis and might be appropriate for resolving other-systems-related matters in healthcare sector. Also, it can aid as a respected/valuable decision-support tool for the doctors, providing further info and visions to help in their diagnosis and treatment strategies.</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

Machine learning technique for Alzheimer’s disease diagnosing using online handwriting data

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • Musatafa Abbas Abbood Albadr,
  • Fahad Taha AL-Dhief,
  • Raad Z. Homod

摘要

The utilization of Data Mining (DM) and Machine Learning (ML) approaches in Alzheimer Disease (AD) diagnosis has lately gained a lot of attention. However, utmost of these studies still required improvement since either they were evaluated utilizing insufficient assessment-measurements; or they were not statistically-evaluated; or both. Recently, the PSO-ELM (Particle Swarm Optimization-Extreme Learning Machine) is considered one-of-the-most well-known and effective ML methods, it has seen as a reputable and efficient technique for categorizing-data, however it hasn’t been employed in AD diagnosis issue. Thus, this study proposes the PSO-ELM model in-order-to lift the accuracy rate of AD diagnosis. The PSO-ELM model has the capability to (i) ability to be utilized on both (binary/multi-class) classification, (ii) prevent overfitting; and also (iii) it has a similar capability to kernel-based SVM (Support Vector Machine) and functions with a structure of neural-network. In this paper, the DARWIN (Diagnosis AlzheimeR WIth haNdwriting) dataset was utilized to assess the PSO-ELM model performance. The experimentations results have demonstrated the distinct performance of the proposed PSO-ELM model, which reached an average of F-Measure 98.58%, recall 100.00%, specificity 97.60%, precision 97.59%, G-Mean 98.60%, MCC 97.22%, and accuracy 98.57%. This designates that the PSO-ELM model is a reliable method for AD diagnosis and might be appropriate for resolving other-systems-related matters in healthcare sector. Also, it can aid as a respected/valuable decision-support tool for the doctors, providing further info and visions to help in their diagnosis and treatment strategies.