Scholarlink

<p>The prevalence of thyroid disease (TD) and thyroid cancer (TC) has steadily increased over the past four decades. Diagnosis typically relies on blood tests interpreted by medical experts, which can delay treatment. Additionally, noisy and imbalanced data often hinder the performance of automated diagnostic systems, resulting in many undiagnosed or untreated cases. This study aims to evaluate the effectiveness of machine learning (ML) and deep learning (DL) algorithms in the early detection and classification of thyroid disorders. This study comprehensively reviews the innovations and challenges of ML and DL approaches, focusing on integrating advanced AI and ML paradigms such as federated learning, explainable AI, quantum ML, distributed intelligence, and reinforcement learning to enhance adaptability and performance. In addition, the review also compares performance metrics such as accuracy, precision, recall, and F1-score. It also analyzes dataset types, preprocessing techniques, and model optimization strategies for thyroid disease prediction. Findings show that algorithms such as Random Forest, Support Vector Machines, and neural networks are commonly used, with varying levels of success. Data quality, balancing methods, and feature selection significantly influence model performance. Despite promising results, many models struggle with generalizability and reliability. This paper envisions an AI leveraging solution for thyroid disease: advances and rationale by addressing these challenges and exploring cutting-edge strategies. AI-based models offer potential for improved early diagnosis of thyroid disorders. The study also highlights the systematic review, pointing out the advantages and disadvantages of ML and DL, which can serve as a roadmap for future research efforts.</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

AI leveraging solution for thyroid disease: advances and rationale

  • 期刊论文

  • 发表日期：

    2025年8月15日

查看全文
  • Shanu Verma,
  • Rashmi Popli,
  • Harish Kumar,
  • Brijesh Kumar Chaurasia

摘要

The prevalence of thyroid disease (TD) and thyroid cancer (TC) has steadily increased over the past four decades. Diagnosis typically relies on blood tests interpreted by medical experts, which can delay treatment. Additionally, noisy and imbalanced data often hinder the performance of automated diagnostic systems, resulting in many undiagnosed or untreated cases. This study aims to evaluate the effectiveness of machine learning (ML) and deep learning (DL) algorithms in the early detection and classification of thyroid disorders. This study comprehensively reviews the innovations and challenges of ML and DL approaches, focusing on integrating advanced AI and ML paradigms such as federated learning, explainable AI, quantum ML, distributed intelligence, and reinforcement learning to enhance adaptability and performance. In addition, the review also compares performance metrics such as accuracy, precision, recall, and F1-score. It also analyzes dataset types, preprocessing techniques, and model optimization strategies for thyroid disease prediction. Findings show that algorithms such as Random Forest, Support Vector Machines, and neural networks are commonly used, with varying levels of success. Data quality, balancing methods, and feature selection significantly influence model performance. Despite promising results, many models struggle with generalizability and reliability. This paper envisions an AI leveraging solution for thyroid disease: advances and rationale by addressing these challenges and exploring cutting-edge strategies. AI-based models offer potential for improved early diagnosis of thyroid disorders. The study also highlights the systematic review, pointing out the advantages and disadvantages of ML and DL, which can serve as a roadmap for future research efforts.