Scholarlink

The thyroid gland is an organ that creates the most significant hormones that control how other parts of the body operate. Hence, early diagnosis of thyroid disease is crucial for proper treatment. This leads to a need for tools to aid understanding and inform medical decisions. This paper presents a novel deep learning model based on Convolutional Neural Networks (CNNs) and Particle Swarm Optimization (PSO) for the early stage diagnosis of thyroid disease. The proposed model aims to classify patients into two categories: Thyroid disease positive and negative. Using the University of California Irvine machine learning repository’s thyroid dataset, we preprocess the data, employing techniques such as transformation, noise removal, sampling, and Min–Max normalization. After preprocessing, categorical data is converted into numerical representations using LabelEncoder from scikit-learn. Next, the proposed approach is compared with various machine learning models, including Random Forest, Logistic Regression, K-Nearest Neighbors (KNN), and AdaBoost, to prove its superior performance of 99.58% accuracy. Additionally, PSO is utilized for feature extraction, identifying crucial features like “sick,” “thyroid surgery,” “TSH,” and “TT4.” This significant improvement demonstrates the potential of deep learning and PSO in enhancing the accuracy and effectiveness of thyroid disease diagnosis. Our new knowledge may help medical professionals identify thyroid issues more precisely and provide patients with better options for care.

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

Early Prediction of Thyroid Disease Through the Integration of Machine Learning Algorithms and Particle Swarm Optimization-Based Feature Extraction: A Novel Approach for Improved Diagnosis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Jemimah Maria Kuruvilla,
  • S. Jeba Priya,
  • G. Jaspher W. Kathrine,
  • A. Shamila Ebenezer

摘要

The thyroid gland is an organ that creates the most significant hormones that control how other parts of the body operate. Hence, early diagnosis of thyroid disease is crucial for proper treatment. This leads to a need for tools to aid understanding and inform medical decisions. This paper presents a novel deep learning model based on Convolutional Neural Networks (CNNs) and Particle Swarm Optimization (PSO) for the early stage diagnosis of thyroid disease. The proposed model aims to classify patients into two categories: Thyroid disease positive and negative. Using the University of California Irvine machine learning repository’s thyroid dataset, we preprocess the data, employing techniques such as transformation, noise removal, sampling, and Min–Max normalization. After preprocessing, categorical data is converted into numerical representations using LabelEncoder from scikit-learn. Next, the proposed approach is compared with various machine learning models, including Random Forest, Logistic Regression, K-Nearest Neighbors (KNN), and AdaBoost, to prove its superior performance of 99.58% accuracy. Additionally, PSO is utilized for feature extraction, identifying crucial features like “sick,” “thyroid surgery,” “TSH,” and “TT4.” This significant improvement demonstrates the potential of deep learning and PSO in enhancing the accuracy and effectiveness of thyroid disease diagnosis. Our new knowledge may help medical professionals identify thyroid issues more precisely and provide patients with better options for care.