Scholarlink

<p>Chronic​‍​‌‍​‍‌​‍​‌‍​‍‌ kidney disease is very common and its negative impact on the quality of life of the afflicted makes it a considerable health problem worldwide. In due course, early detection and accurate diagnosis of CKD will make the first intervention that will actually ameliorate the management of the disease possible. Nevertheless, it is equally true that the models developed for the prediction of CKD have been beset with several problems only one of which is their capacity to represent limited features, their incapability to handle complex data trends, and their incapability to recognize different levels of the disease. Therefore, the paper proposes a new algorithm DenseSwinXNet that will be an effective hybrid framework consisting of DenseNet, Swin Transformer, and MobileNet components. It gets the feature of dense connectivity in DenseNet to strong feature learning, the highly advanced self-attention features of the ATTnet family and Swin Transformer of global and local contextual information learning, and the computational efficiency of convolutional operation in models like MobileNet. These features empower DSXNet to have control over various forms of data as well as advanced forms of data patterns which are numerous in CKD data. The new model is breaking new ground in the sharp planning of CKD in many ways. Firstly, it comprises the dense blocks used by DenseNet to ensure maximum extraction and gradient flow, which will help tremendously in propelling the model in learning complex patterns. The main innovation in DSXNet is the fused framework as it merges the intricate methods of dense connectivity, attention mechanisms, and efficient convolutional mechanisms that were designed to alleviate the disadvantages of the previous models. By means of these three ways, DSXNet might accomplish the great benefits of the prediction in both accuracy and speed. The findings showed that DSXNet outperformed the advanced and traditional methods in all the metrics of precision, recall, F1-score, and AUROC for CKD detection and management, thus, setting a new ​‍​‌‍​‍‌​‍​‌‍​‍‌standard.</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

DSXNet: A Cutting-Edge Hybrid Model for Accurate Chronic Kidney Disease Detection and Classification

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月28日

查看全文
  • R Kalaivani,
  • A Bharathi

摘要

Chronic​‍​‌‍​‍‌​‍​‌‍​‍‌ kidney disease is very common and its negative impact on the quality of life of the afflicted makes it a considerable health problem worldwide. In due course, early detection and accurate diagnosis of CKD will make the first intervention that will actually ameliorate the management of the disease possible. Nevertheless, it is equally true that the models developed for the prediction of CKD have been beset with several problems only one of which is their capacity to represent limited features, their incapability to handle complex data trends, and their incapability to recognize different levels of the disease. Therefore, the paper proposes a new algorithm DenseSwinXNet that will be an effective hybrid framework consisting of DenseNet, Swin Transformer, and MobileNet components. It gets the feature of dense connectivity in DenseNet to strong feature learning, the highly advanced self-attention features of the ATTnet family and Swin Transformer of global and local contextual information learning, and the computational efficiency of convolutional operation in models like MobileNet. These features empower DSXNet to have control over various forms of data as well as advanced forms of data patterns which are numerous in CKD data. The new model is breaking new ground in the sharp planning of CKD in many ways. Firstly, it comprises the dense blocks used by DenseNet to ensure maximum extraction and gradient flow, which will help tremendously in propelling the model in learning complex patterns. The main innovation in DSXNet is the fused framework as it merges the intricate methods of dense connectivity, attention mechanisms, and efficient convolutional mechanisms that were designed to alleviate the disadvantages of the previous models. By means of these three ways, DSXNet might accomplish the great benefits of the prediction in both accuracy and speed. The findings showed that DSXNet outperformed the advanced and traditional methods in all the metrics of precision, recall, F1-score, and AUROC for CKD detection and management, thus, setting a new ​‍​‌‍​‍‌​‍​‌‍​‍‌standard.