Home Mathematics Hybrid Attention-Driven Network for Predictive Healthcare Using Machine Learning and Data Analytics Perspective
Chapter
Licensed
Unlicensed Requires Authentication

Hybrid Attention-Driven Network for Predictive Healthcare Using Machine Learning and Data Analytics Perspective

  • Shakila Basheer
Become an author with De Gruyter Brill
Explore this Subject How to publish with us
Machine Learning in Healthcare
This chapter is in the book Machine Learning in Healthcare

Abstract

The integration of machine learning (ML) and healthcare analytics has opened new previously unseen doors for risk assessment, early diagnosis, and personalized planning of therapy. With the growing expansion of medical imaging, electronic health records (EHRs), and physiological signals, today it is both possible and imperative to use data-driven approaches to uncover meaningful insights. (i) Limited generalizability across patient populations due to insufficient or skewed data; (ii) interpretability of black-box models, undermining clinical trust; and (iii) inadequate multimodal integration of data sources, resulting in isolated decision-making, are the three key issues with current ML-based frameworks, despite spectacular advancements. To overcome these, we present HAD-Net (hybrid attention-driven network), a multimodal deep learning architecture that combines structured EHR data with unstructured imaging modalities through the combined strength of self-attention mechanisms, graph neural networks (GNNs), and bidirectional long short-term memory (BiLSTM). A Bayesian-tuned adaptive learning scheduler is utilized to adaptively tune the model for better convergence, and SHAP (SHapley Additive exPlanations) is added to provide understandable risk predictions. A dataset of the NIH ChestX-ray14 collection and the MIMIC-III clinical database was used to test the proposed HAD-Net. With an AUROC of 94.3%, accuracy of 91.7%, and F1-score of 92.1%, the HAD-Net outperformed the best-performing baseline by 9.5% and 12.4%, respectively, over conventional machine learning models (random forest and logistic regression) and novel deep models (CNN-LSTM and transformer-based models). Moreover, the addition of SHAP explanations enhanced interpretability of the model and enabled biomarker discovery for clinically relevant biomarkers like abnormal PaO₂/FiO₂ ratios, raised troponin, and radiographic infiltrates. These findings validate the applicability of HAD-Net for effective, multimodal analytics to provide personalized and predictive care, which in turn enables proactive clinical decision-making and minimizes avoidable adverse consequences.

Abstract

The integration of machine learning (ML) and healthcare analytics has opened new previously unseen doors for risk assessment, early diagnosis, and personalized planning of therapy. With the growing expansion of medical imaging, electronic health records (EHRs), and physiological signals, today it is both possible and imperative to use data-driven approaches to uncover meaningful insights. (i) Limited generalizability across patient populations due to insufficient or skewed data; (ii) interpretability of black-box models, undermining clinical trust; and (iii) inadequate multimodal integration of data sources, resulting in isolated decision-making, are the three key issues with current ML-based frameworks, despite spectacular advancements. To overcome these, we present HAD-Net (hybrid attention-driven network), a multimodal deep learning architecture that combines structured EHR data with unstructured imaging modalities through the combined strength of self-attention mechanisms, graph neural networks (GNNs), and bidirectional long short-term memory (BiLSTM). A Bayesian-tuned adaptive learning scheduler is utilized to adaptively tune the model for better convergence, and SHAP (SHapley Additive exPlanations) is added to provide understandable risk predictions. A dataset of the NIH ChestX-ray14 collection and the MIMIC-III clinical database was used to test the proposed HAD-Net. With an AUROC of 94.3%, accuracy of 91.7%, and F1-score of 92.1%, the HAD-Net outperformed the best-performing baseline by 9.5% and 12.4%, respectively, over conventional machine learning models (random forest and logistic regression) and novel deep models (CNN-LSTM and transformer-based models). Moreover, the addition of SHAP explanations enhanced interpretability of the model and enabled biomarker discovery for clinically relevant biomarkers like abnormal PaO₂/FiO₂ ratios, raised troponin, and radiographic infiltrates. These findings validate the applicability of HAD-Net for effective, multimodal analytics to provide personalized and predictive care, which in turn enables proactive clinical decision-making and minimizes avoidable adverse consequences.

Chapters in this book

  1. Frontmatter I
  2. Contents V
  3. Early Prediction of Chronic Kidney Disease Using a Novel Hybrid Regularized Adaptive Boosting Algorithm: An Advanced Machine Learning Approach 1
  4. DigiCure: A Patient-Centric Framework for Digital Transformation in Healthcare 21
  5. Exploring Machine Learning Approaches for Maximizing the Likelihood of Diabetes Classification 41
  6. A Hybrid Machine Learning Model for Risk Stratification and Functional Outcome Prediction in Stroke Survivors 61
  7. Data-Driven Machine Learning Strategies for Oncological Disease Prediction and Early-Stage Detection 83
  8. Machine Learning Applications in Mental Health: Ensemble-Based Predictive Modeling for Depression and Anxiety detection 103
  9. Privacy-Preserving Machine Learning in Clinical Research: Using Federated Learning to Protect Patient Data 129
  10. EpiCastNet: A Spatiotemporal Hybrid Learning Framework for Real-Time Epidemic Forecasting 149
  11. Machine Learning for Early Detection of Chronic Diseases: A Case Study in Diabetes Prediction 171
  12. Machine Learning Techniques for Healthcare 193
  13. Applications and Benefits of Machine Learning in Healthcare 215
  14. Intelligent Treatment Recommendation Using CareRecNet: A Patient-Centered Approach to Digital Health Transformation 233
  15. Reinforcement-Driven Graph Neural Framework for Personalized and Proactive Patient Care in Digital Health Systems 251
  16. Hybrid Attention-Driven Network for Predictive Healthcare Using Machine Learning and Data Analytics Perspective 271
  17. MSAG-DFE: A Multi-scale Attention-Guided Deep Feature Extraction Framework for Enhanced Medical Image Diagnostics 287
  18. On Mental Health Monitoring Using Commercial Wearable Devices and Machine Intelligence 305
  19. Enhancing Healthcare Delivery Through Evidence-Based Data Utilization 335
  20. AGBO-CP: An Adaptive Gradient Boosted Optimization Framework for Enhanced Clinical Prediction Accuracy 367
  21. A Hierarchical Cross-Fusion Feature Extraction Network for Accurate Cervical Cancer Classification Using Cytology Images 387
  22. Analyzing the Impact of Social Network on Epidemiological Spread in the Healthcare Sector 409
  23. Intelligent Interventions: Practical Applications of Machine Learning for Data-Driven Decision-Making in Healthcare 431
  24. Stress Recognition Through Physiological and Behavioral Signals: A Machine Learning Perspective 453
  25. MediChain-FL: A Federated Blockchain Framework for Privacy-Preserving and Intelligent Healthcare Data Exchange 485
  26. Reinforced Multi-objective Optimization Framework for Adaptive Healthcare Decision Intelligence 503
  27. Index
https://doi.org/10.1515/9783111541235-014

Chapters in this book

  1. Frontmatter I
  2. Contents V
  3. Early Prediction of Chronic Kidney Disease Using a Novel Hybrid Regularized Adaptive Boosting Algorithm: An Advanced Machine Learning Approach 1
  4. DigiCure: A Patient-Centric Framework for Digital Transformation in Healthcare 21
  5. Exploring Machine Learning Approaches for Maximizing the Likelihood of Diabetes Classification 41
  6. A Hybrid Machine Learning Model for Risk Stratification and Functional Outcome Prediction in Stroke Survivors 61
  7. Data-Driven Machine Learning Strategies for Oncological Disease Prediction and Early-Stage Detection 83
  8. Machine Learning Applications in Mental Health: Ensemble-Based Predictive Modeling for Depression and Anxiety detection 103
  9. Privacy-Preserving Machine Learning in Clinical Research: Using Federated Learning to Protect Patient Data 129
  10. EpiCastNet: A Spatiotemporal Hybrid Learning Framework for Real-Time Epidemic Forecasting 149
  11. Machine Learning for Early Detection of Chronic Diseases: A Case Study in Diabetes Prediction 171
  12. Machine Learning Techniques for Healthcare 193
  13. Applications and Benefits of Machine Learning in Healthcare 215
  14. Intelligent Treatment Recommendation Using CareRecNet: A Patient-Centered Approach to Digital Health Transformation 233
  15. Reinforcement-Driven Graph Neural Framework for Personalized and Proactive Patient Care in Digital Health Systems 251
  16. Hybrid Attention-Driven Network for Predictive Healthcare Using Machine Learning and Data Analytics Perspective 271
  17. MSAG-DFE: A Multi-scale Attention-Guided Deep Feature Extraction Framework for Enhanced Medical Image Diagnostics 287
  18. On Mental Health Monitoring Using Commercial Wearable Devices and Machine Intelligence 305
  19. Enhancing Healthcare Delivery Through Evidence-Based Data Utilization 335
  20. AGBO-CP: An Adaptive Gradient Boosted Optimization Framework for Enhanced Clinical Prediction Accuracy 367
  21. A Hierarchical Cross-Fusion Feature Extraction Network for Accurate Cervical Cancer Classification Using Cytology Images 387
  22. Analyzing the Impact of Social Network on Epidemiological Spread in the Healthcare Sector 409
  23. Intelligent Interventions: Practical Applications of Machine Learning for Data-Driven Decision-Making in Healthcare 431
  24. Stress Recognition Through Physiological and Behavioral Signals: A Machine Learning Perspective 453
  25. MediChain-FL: A Federated Blockchain Framework for Privacy-Preserving and Intelligent Healthcare Data Exchange 485
  26. Reinforced Multi-objective Optimization Framework for Adaptive Healthcare Decision Intelligence 503
  27. Index
Sign up now to receive a 20% welcome discount
Institutional Access
How does access work?
Have an idea on how to improve our website?
Please write us.
© 2025 De Gruyter Brill
Downloaded on 27.12.2025 from https://www.degruyterbrill.com/document/doi/10.1515/9783111541235-014/html
Scroll to top button