Scholarlink

<p>Biliary atresia (BA) is a rare and devastating bile duct disease that occurs during the neonatal period. Timely identification and prompt surgical intervention is critical for improving the outcome. The aim of the study is to propose a hybrid machine learning (ML) based prediction model for the detection of BA. The framework makes use of ASO-XGB (adaptive selective orthogonality-extreme gradient boost) which integrates ASO into ResNet-50 architecture to ensure features extraction from the disease dataset. The XGBoost algorithm will classify the outcomes of disease once it takes deep features from the neural network. ASO uses orthogonality regularization selectively on layers with low gradient norms through its application to maintain stable deep feature learning. The extracted features move from XGBoost for classification, which builds model generalization with overfitting prevention. The results show that the proposed ASO-XGB attains the highest accuracy (97.21%) thereby surpassing existing studies and techniques. Thus, ASO-XGB establishes a successful equilibrium between selecting relevant features and running accurate classifications and operation time, establishing it as an advanced AI-based clinical decision tool.</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

ASO-XGB: an adaptive selective orthogonality and XGBoost-based hybrid model for detection of biliary atresia

  • 期刊论文

  • 发表日期：

    2025年12月27日

查看全文
  • Kumari Monika,
  • Kushal Kanwar

摘要

Biliary atresia (BA) is a rare and devastating bile duct disease that occurs during the neonatal period. Timely identification and prompt surgical intervention is critical for improving the outcome. The aim of the study is to propose a hybrid machine learning (ML) based prediction model for the detection of BA. The framework makes use of ASO-XGB (adaptive selective orthogonality-extreme gradient boost) which integrates ASO into ResNet-50 architecture to ensure features extraction from the disease dataset. The XGBoost algorithm will classify the outcomes of disease once it takes deep features from the neural network. ASO uses orthogonality regularization selectively on layers with low gradient norms through its application to maintain stable deep feature learning. The extracted features move from XGBoost for classification, which builds model generalization with overfitting prevention. The results show that the proposed ASO-XGB attains the highest accuracy (97.21%) thereby surpassing existing studies and techniques. Thus, ASO-XGB establishes a successful equilibrium between selecting relevant features and running accurate classifications and operation time, establishing it as an advanced AI-based clinical decision tool.