Scholarlink

<p>Orthopedic disease classification refers to a systematic approach used to categorize and describe different musculoskeletal disorders that affect bones, joints, muscles, ligaments, tendons, and related anatomical structures. This structured framework is essential for achieving accurate diagnoses, developing effective treatment strategies, and maintaining efficient communication among medical practitioners. Explainable Artificial Intelligence (XAI) plays a key role in enhancing the transparency of classification models by enabling them to not only provide accurate predictions but also explain the reasoning behind their outcomes. Conventional machine learning methods, while powerful, often act as “black boxes,” making it difficult to interpret the logic behind their decisions. To overcome this limitation, this study introduces an Explainable Artificial Intelligence-Medical (XAI-Med) model specifically designed for the classification of orthopedic diseases. The proposed XAI-Med framework integrates Adaptive Boosting (AdaBoost) to strengthen predictive performance and SHapley Additive exPlanations (SHAP) to deliver interpretable insights into feature contributions. Moreover, a newly developed Binary Depth First Search (BDFS) algorithm is employed for feature selection to identify the most relevant attributes influencing classification results. This hybrid design aims to provide meaningful interpretability while improving diagnostic decision-making in orthopedic disease classification. The performance of XAI-Med is systematically compared with several traditional machine learning algorithms, including Logistic Regression (LR), Support Vector Machine (SVM), Ridge Classifier (RC), Gradient Boosting (GB), Random Forest (RF), K-Nearest Neighbor (KNN), and Decision Tree (DT). A range of evaluation measures, including accuracy, sensitivity, specificity, balanced accuracy, F1-score, and AUC were applied to assess model performance. The experimental findings show that the XAI-Med approach delivers superior results compared with the benchmark techniques, recording a leading accuracy of 0.971.</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 orthopedic disease classification based on explainable artificial intelligence and binary depth first search algorithm

  • 期刊论文

  • 发表日期：

    2026年3月26日

查看全文
  • Ahmed M. Elshewey,
  • Amel Ali Alhussan,
  • Doaa Sami Khafaga,
  • El-Sayed M. Elkenawy,
  • Zahraa Tarek

摘要

Orthopedic disease classification refers to a systematic approach used to categorize and describe different musculoskeletal disorders that affect bones, joints, muscles, ligaments, tendons, and related anatomical structures. This structured framework is essential for achieving accurate diagnoses, developing effective treatment strategies, and maintaining efficient communication among medical practitioners. Explainable Artificial Intelligence (XAI) plays a key role in enhancing the transparency of classification models by enabling them to not only provide accurate predictions but also explain the reasoning behind their outcomes. Conventional machine learning methods, while powerful, often act as “black boxes,” making it difficult to interpret the logic behind their decisions. To overcome this limitation, this study introduces an Explainable Artificial Intelligence-Medical (XAI-Med) model specifically designed for the classification of orthopedic diseases. The proposed XAI-Med framework integrates Adaptive Boosting (AdaBoost) to strengthen predictive performance and SHapley Additive exPlanations (SHAP) to deliver interpretable insights into feature contributions. Moreover, a newly developed Binary Depth First Search (BDFS) algorithm is employed for feature selection to identify the most relevant attributes influencing classification results. This hybrid design aims to provide meaningful interpretability while improving diagnostic decision-making in orthopedic disease classification. The performance of XAI-Med is systematically compared with several traditional machine learning algorithms, including Logistic Regression (LR), Support Vector Machine (SVM), Ridge Classifier (RC), Gradient Boosting (GB), Random Forest (RF), K-Nearest Neighbor (KNN), and Decision Tree (DT). A range of evaluation measures, including accuracy, sensitivity, specificity, balanced accuracy, F1-score, and AUC were applied to assess model performance. The experimental findings show that the XAI-Med approach delivers superior results compared with the benchmark techniques, recording a leading accuracy of 0.971.