Scholarlink

<p>The prevalence and mortality of chronic diseases are rising globally. As a result, early diagnosis has emerged as a key area of study to improve patient survival. Numerous studies have documented classification strategies for the prediction of particular diseases. However, their inefficiency in feature extraction has had a detrimental impact on the prediction rate. The current study aims to predict seven primary chronic diseases that impact individuals globally: diabetes, liver, thyroid, heart disease, kidney, Alzheimer, and breast cancer. This will help address concerns about the poor accuracy rate. Therefore, in this research, we propose a novel fuzzy-based deep learning and AI-based research framework for multiple chronic disease prediction. Initially, some preprocessing phases are used to standardize the data. The SMOTE approach is used to tackle the issue of class imbalance. After the completion of the preprocessing phase, the essential features are extracted. The LeViT approach and the SincNet approach are ensembles used to extract the features. Using the Fuzzy Pooling Convolutional Neural Network (FPCNN), multiple chronic diseases are predicted. To monitor the health records of chronic patients, the AI-based Scrutiny Boosted Graph Convolutional LSTM (SBGC-LSTM) is utilized. Finally, to show the impact of the prediction outcome, Shapley Additive Explanations is employed. Our proposed method had the best accuracy, at 99.28%. Nonetheless, it is discovered that the characteristics of the data used for classification affect how well the classification model performs. Our findings outperform alternative benchmark approaches when compared to different chronic disease datasets. Furthermore, it is demonstrated that our suggested processing method takes less time than methods based on Support Vector Machines (SVM), deep learning, and Random Forests (RF). Through the creation of online diagnosis tools, our research may contribute to the early diagnosis of chronic illnesses in hospitals.</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

An Explainable Fuzzy Pooling CNN Model for Multi-chronic Disease Prediction with Scrutiny Boosted Graph LSTM Monitoring

  • 期刊论文

  • 发表日期：

    2026年5月7日

查看全文
  • Bondala Anil Kumar,
  • Lella Kranthi Kumar

摘要

The prevalence and mortality of chronic diseases are rising globally. As a result, early diagnosis has emerged as a key area of study to improve patient survival. Numerous studies have documented classification strategies for the prediction of particular diseases. However, their inefficiency in feature extraction has had a detrimental impact on the prediction rate. The current study aims to predict seven primary chronic diseases that impact individuals globally: diabetes, liver, thyroid, heart disease, kidney, Alzheimer, and breast cancer. This will help address concerns about the poor accuracy rate. Therefore, in this research, we propose a novel fuzzy-based deep learning and AI-based research framework for multiple chronic disease prediction. Initially, some preprocessing phases are used to standardize the data. The SMOTE approach is used to tackle the issue of class imbalance. After the completion of the preprocessing phase, the essential features are extracted. The LeViT approach and the SincNet approach are ensembles used to extract the features. Using the Fuzzy Pooling Convolutional Neural Network (FPCNN), multiple chronic diseases are predicted. To monitor the health records of chronic patients, the AI-based Scrutiny Boosted Graph Convolutional LSTM (SBGC-LSTM) is utilized. Finally, to show the impact of the prediction outcome, Shapley Additive Explanations is employed. Our proposed method had the best accuracy, at 99.28%. Nonetheless, it is discovered that the characteristics of the data used for classification affect how well the classification model performs. Our findings outperform alternative benchmark approaches when compared to different chronic disease datasets. Furthermore, it is demonstrated that our suggested processing method takes less time than methods based on Support Vector Machines (SVM), deep learning, and Random Forests (RF). Through the creation of online diagnosis tools, our research may contribute to the early diagnosis of chronic illnesses in hospitals.