Scholarlink

<p>Cardiovascular diseases (CVD) have consistently ranked among the leading causes of mortality worldwide, accounting for a significant proportion of deaths across all age groups and regions. These diseases encompass a range of conditions affecting the heart and blood vessels, including coronary artery disease, heart failure, stroke, and hypertension. With the notion of timely intervention and risk management, a hybrid deep learning approach incorporating Bidirectional Gated Recurrent Unit (Bi-GRU) and extreme Gradient Boosting (XGBoost) for precise CVD prediction has been proposed. In addition, leveraging Internet-of-Things (IoT) devices in heart disease detection enables self-assessments to monitor for irregularities in heart function which aids in the early detection of heart disease. Several IoT devices including Electrocardiogram (ECG) sensors, glucometers for measuring blood pressure and pulse are employed. The proposed system leverages real-time model testing using IoT-enabled wearable devices to analyze the vital signs as well as other health evaluation parameters. This information is processed and monitored using the proposed model containing ensembled deep learning and machine learning algorithms, specifically trained to identify the risk of presence of cardiovascular diseases. Furthermore, to maintain reliability of the model’s predictions, we incorporate explainable artificial intelligence techniques, which provide transparent and interpretable illustrations helping the decision-making process of these algorithms. The trained model is evaluated using the Mendeley Cardiovascular Disease dataset along with real-time data collected from various IoT sensors monitoring patients. This approach tests and validates the efficacy of the proposed hybrid deep learning model, which achieves an accuracy of 98%. The proposed approach seeks to integrate predictive analytics with real-time IoT applications in the medical field, enhancing cardiovascular care through improved outcomes and more informed decision-making.</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

Empowering cardiovascular disease detection with IoT-enabled interpretable deep learning driven approach using explainable AI

  • 期刊论文

  • 发表日期：

    2026年7月21日

查看全文
  • Rohit Ahuja,
  • Amisha Jain,
  • Muskan,
  • Shashank Sheshar Singh

摘要

Cardiovascular diseases (CVD) have consistently ranked among the leading causes of mortality worldwide, accounting for a significant proportion of deaths across all age groups and regions. These diseases encompass a range of conditions affecting the heart and blood vessels, including coronary artery disease, heart failure, stroke, and hypertension. With the notion of timely intervention and risk management, a hybrid deep learning approach incorporating Bidirectional Gated Recurrent Unit (Bi-GRU) and extreme Gradient Boosting (XGBoost) for precise CVD prediction has been proposed. In addition, leveraging Internet-of-Things (IoT) devices in heart disease detection enables self-assessments to monitor for irregularities in heart function which aids in the early detection of heart disease. Several IoT devices including Electrocardiogram (ECG) sensors, glucometers for measuring blood pressure and pulse are employed. The proposed system leverages real-time model testing using IoT-enabled wearable devices to analyze the vital signs as well as other health evaluation parameters. This information is processed and monitored using the proposed model containing ensembled deep learning and machine learning algorithms, specifically trained to identify the risk of presence of cardiovascular diseases. Furthermore, to maintain reliability of the model’s predictions, we incorporate explainable artificial intelligence techniques, which provide transparent and interpretable illustrations helping the decision-making process of these algorithms. The trained model is evaluated using the Mendeley Cardiovascular Disease dataset along with real-time data collected from various IoT sensors monitoring patients. This approach tests and validates the efficacy of the proposed hybrid deep learning model, which achieves an accuracy of 98%. The proposed approach seeks to integrate predictive analytics with real-time IoT applications in the medical field, enhancing cardiovascular care through improved outcomes and more informed decision-making.