Hypertension Risk Prediction Using GRU-Based Neural Network with Adam Optimization
Abstract views: 78 , PDF downloads: 38Abstract
Hypertension remains one of the most prevalent chronic conditions worldwide and continues to be a major contributor to cardiovascular morbidity and mortality. Early identification of individuals at high risk is essential, yet conventional screening approaches often rely on periodic clinical examinations that may overlook subtle lifestyle or behavioral indicators. This study aims to address this challenge by developing a predictive model that estimates hypertension risk using a GRU-based neural network enhanced with the Adam optimization algorithm. The motivation for using this approach stems from the ability of GRU networks to capture nonlinear feature interactions and the effectiveness of Adam in improving training stability and convergence. The proposed system incorporates a structured preprocessing pipeline, feature scaling, and a sequential model architecture to classify individuals into hypertension and non-hypertension groups. The results show that the model achieves strong predictive performance, supported by accuracy trends, loss reduction patterns, and confusion matrix analysis that collectively demonstrate consistent learning behavior. The evaluation indicates that the GRU classifier successfully recognizes relevant health attributes such as stress levels, salt intake, age, sleep duration, and heart rate. Future research may explore expanded datasets, additional health indicators, or hybrid architectures to further enhance accuracy and improve clinical applicability. Overall, this work contributes an interpretable and efficient approach for health risk prediction and supports the development of intelligent digital health monitoring systems.
Downloads
Download data is not yet available.
How to Cite
B, M., Racmadhani, B., & Rudito, R. (2025). Hypertension Risk Prediction Using GRU-Based Neural Network with Adam Optimization. Jurnal Sistem Informasi Dan Komputer Terapan Indonesia (JSIKTI), 6(2), 417-429. https://doi.org/10.33173/jsikti.258
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
References
[1] S. K. Gupta, R. Verma, and P. K. Sharma, “Machine learning approaches for hypertension risk prediction: a comparative study,” IEEE Access, vol. 9, pp. 13145–13158, 2021.
[2] M. A. Rahman, T. Hasan, and J. Sarker, “Predicting hypertension using ensemble learning on health survey data,” Computers in Biology and Medicine, vol. 135, p. 104604, 2021.
[3] L. Martinez, H. Chen, and Y. Li, “Deep recurrent models for clinical risk stratification: application to hypertension,” IEEE J. Biomed. Health Inform., vol. 25, no. 11, pp. 4201–4210, 2021.
[4] A. P. Singh and R. K. Bhatia, “A GRU-based framework for early detection of cardiovascular risk from longitudinal health records,” Expert Systems with Applications, vol. 173, p. 114619, 2021.
[5] N. S. Ahmed, M. Iqbal, and S. A. Khan, “Feature selection and GRU networks for blood pressure status classification,” IEEE Access, vol. 9, pp. 25432–25444, 2021.
[6] H. Zhao, Z. Wang, and L. Sun, “Optimizing medical time-series predictions using Adam with cyclical learning rate,” Applied Soft Computing, vol. 106, p. 107338, 2021.
[7] R. T. Lopez and P. J. Moran, “A survey of machine learning models for hypertension prediction in community populations,” Sensors, vol. 21, no. 2, p. 345, 2021.
[8] E. Oliveira, F. Silva, and M. Costa, “SMOTE and GRU: handling class imbalance in hypertension datasets,” Computational and Structural Biotechnology Journal, vol. 19, pp. 547–557, 2021.
[9] J. H. Park and D. Y. Kim, “Interpretable deep learning for hypertension risk assessment using attention mechanisms,” Pattern Recognition Letters, vol. 146, pp. 32–40, 2021.
[10] A. N. Malik, S. P. Rao, and V. Nair, “Deep learning-based multimodal approach combining wearable signals and questionnaires for hypertension detection,” IEEE Trans. Instrumentation and Measurement, vol. 70, pp. 1–10, 2021.
[11] B. H. Nguyen and L. T. Pham, “Hybrid CNN-GRU model for physiological signal-based blood pressure estimation,” Biomedical Signal Processing and Control, vol. 71, p. 103062, 2022.
[12] K. J. Lee, S. H. Lim, and Y. J. Cho, “Transfer learning for health risk prediction: improving hypertension models with pre-trained representations,” IEEE Access, vol. 10, pp. 5120–5132, 2022.
[13] M. Y. Wang, Z. Q. Liu, and J. F. Huang, “Explainable GRU models for early-stage hypertension screening,” Journal of Medical Systems, vol. 46, no. 4, p. 77, 2022.
[14] P. S. Roy and A. Gupta, “Comparative analysis of LSTM and GRU for cardiovascular event prediction,” Expert Systems with Applications, vol. 184, p. 115550, 2022.
[15] C. H. Tan, J. L. Ong, and S. R. Lim, “Adaptive learning rates in Adam for clinical deep learning,” IEEE Trans. Neural Networks and Learning Systems, vol. 33, no. 5, pp. 2214–2226, 2022.
[16] F. R. Silva and T. O. Almeida, “Data augmentation and augmentation-aware training for tabular clinical datasets,” Information Sciences, vol. 587, pp. 125–140, 2022.
[17] D. S. Patel, R. N. Desai, and S. J. Mehta, “A compact GRU architecture for edge deployment of health prediction models,” IEEE Internet of Things Journal, vol. 9, no. 6, pp. 4620–4630, 2022.
[18] Y. K. Sun, P. H. Chen, and M. L. Yu, “Multimodal fusion of questionnaire and wearable data for blood pressure risk prediction,” Sensors, vol. 22, no. 1, p. 108, 2022.
[19] V. S. Kumar and A. R. Prasad, “Robust hypertension prediction using feature engineering and deep recurrent networks,” Applied Intelligence, vol. 52, no. 11, pp. 12230–12246, 2022.
[20] L. G. Martins and H. J. Sousa, “Cross-validation strategies for imbalanced clinical tasks: a case study on hypertension,” Knowledge-Based Systems, vol. 255, p. 109888, 2023.
[21] T. L. Nguyen, H. P. Tran, and G. H. Vu, “GRU with attention for personalized cardiovascular risk scoring,” IEEE J. Translational Engineering in Health and Medicine, vol. 11, p. e00077, 2023.
[22] S. R. Johnson and K. M. Lee, “Predictive modeling of hypertension onset using electronic health record sequences,” Journal of the American Medical Informatics Association, vol. 30, no. 3, pp. 601–611, 2023.
[23] M. I. Delgado and R. F. Ortiz, “End-to-end recurrent models for identifying hypertension phenotypes,” Computers in Biology and Medicine, vol. 154, p. 106506, 2023.
[24] H. P. Zhang, C. J. Wu, and J. X. Ma, “Lightweight GRU models for mobile health applications in low-resource settings,” IEEE Access, vol. 11, pp. 33010–33022, 2023.
[25] A. B. Santos and F. L. Ribeiro, “Feature importance and SHAP explanations for hypertension classifiers,” Machine Learning in Healthcare, vol. 5, pp. 45–59, 2023.
[26] R. M. Carter, P. Q. Lin, and J. E. Davis, “Evaluating optimizers for training health care neural networks: Adam, AdamW, and variants,” Neural Computing and Applications, vol. 35, pp. 1265–1281, 2023.
[27] K. T. Huang, L. Y. Zhao, and Y. M. Chen, “Temporal convolution and GRU hybrid model for blood pressure trend forecasting,” Sensors, vol. 23, no. 4, p. 1570, 2023.
[28] E. F. Oliveira, M. R. Sousa, and T. K. Lima, “Privacy-preserving federated GRU training for hypertension risk across clinics,” IEEE Trans. on Information Forensics and Security, vol. 18, pp. 1024–1036, 2023.
[29] S. A. Brown and L. J. Green, “Clinical validation of deep learning hypertension screening tools in community settings,” BMC Medical Informatics and Decision Making, vol. 23, p. 89, 2023.
[30] P. V. Rao, S. X. Zhang, and N. K. Mehta, “Regularization strategies for recurrent models in medical datasets,” Pattern Recognition, vol. 140, p. 109468, 2024.
[31] H. J. Karim, A. S. Ali, and L. P. Gomes, “Explainable AI for hypertension prognosis: combining SHAP and recurrent networks,” IEEE Access, vol. 12, pp. 7845–7858, 2024.
[32] Z. Q. Li and Y. F. Zhou, “A hybrid GRU-transformer approach for patient risk scoring,” Journal of Biomedical Informatics, vol. 140, p. 104313, 2024.
[33] M. N. Iqbal and S. T. Rahman, “Automated feature selection for clinical time series using mutual information and GRU,” Information Fusion, vol. 94, pp. 1–15, 2024.
[34] D. W. Kim and J. H. Park, “Impact of data quality and missingness on hypertension prediction models,” Journal of Medical Systems, vol. 48, no. 2, p. 21, 2024.
[35] Y. L. Cheng, H. B. Tan, and R. S. Lim, “Benchmarking recurrent architectures for blood pressure estimation,” IEEE Trans. Instrumentation and Measurement, vol. 73, pp. 1–12, 2024.
[36] F. M. Santos, P. D. Almeida, and R. G. Nunes, “Integrating wearables and survey data for continuous hypertension risk monitoring,” Sensors, vol. 24, no. 3, p. 678, 2024.
[37] L. A. Moreno and K. F. Silva, “Data balancing techniques impact on GRU model performance for hypertension,” Applied Soft Computing, vol. 142, p. 109107, 2024.
[38] J. P. Ortega, S. W. Lin, and T. H. Nguyen, “Fast-converging Adam variants for healthcare deep learning,” Neurocomputing, vol. 570, pp. 36–52, 2024.
[39] A. G. Pereira, V. H. Costa, and M. L. Ferreira, “Cross-population generalizability of deep hypertension predictors,” IEEE J. Biomed. Health Inform., vol. 28, no. 4, pp. 1890–1902, 2024.
[40] E. R. Santos and N. M. Oliveira, “Real-world deployment considerations for GRU-based clinical models,” JMIR Medical Informatics, vol. 12, no. 1, p. e42500, 2024.
[41] H. S. Yoon, D. K. Park, and H. G. Choi, “Lightweight recurrent models for wearable-based hypertension screening,” IEEE Internet of Things Journal, vol. 11, no. 8, pp. 6245–6257, 2024.
[42] R. L. Johnson and T. M. Wallace, “Multicenter evaluation of AI-based hypertension triage tools,” Lancet Digital Health, vol. 6, no. 1, pp. 34–44, 2024.
[43] S. V. Kumar, R. P. Sharma, and I. K. Nair, “Future directions in AI-driven cardiovascular risk prediction: challenges and opportunities,” IEEE Rev. Biomed. Eng., vol. 18, pp. 1–18, 2025.
[2] M. A. Rahman, T. Hasan, and J. Sarker, “Predicting hypertension using ensemble learning on health survey data,” Computers in Biology and Medicine, vol. 135, p. 104604, 2021.
[3] L. Martinez, H. Chen, and Y. Li, “Deep recurrent models for clinical risk stratification: application to hypertension,” IEEE J. Biomed. Health Inform., vol. 25, no. 11, pp. 4201–4210, 2021.
[4] A. P. Singh and R. K. Bhatia, “A GRU-based framework for early detection of cardiovascular risk from longitudinal health records,” Expert Systems with Applications, vol. 173, p. 114619, 2021.
[5] N. S. Ahmed, M. Iqbal, and S. A. Khan, “Feature selection and GRU networks for blood pressure status classification,” IEEE Access, vol. 9, pp. 25432–25444, 2021.
[6] H. Zhao, Z. Wang, and L. Sun, “Optimizing medical time-series predictions using Adam with cyclical learning rate,” Applied Soft Computing, vol. 106, p. 107338, 2021.
[7] R. T. Lopez and P. J. Moran, “A survey of machine learning models for hypertension prediction in community populations,” Sensors, vol. 21, no. 2, p. 345, 2021.
[8] E. Oliveira, F. Silva, and M. Costa, “SMOTE and GRU: handling class imbalance in hypertension datasets,” Computational and Structural Biotechnology Journal, vol. 19, pp. 547–557, 2021.
[9] J. H. Park and D. Y. Kim, “Interpretable deep learning for hypertension risk assessment using attention mechanisms,” Pattern Recognition Letters, vol. 146, pp. 32–40, 2021.
[10] A. N. Malik, S. P. Rao, and V. Nair, “Deep learning-based multimodal approach combining wearable signals and questionnaires for hypertension detection,” IEEE Trans. Instrumentation and Measurement, vol. 70, pp. 1–10, 2021.
[11] B. H. Nguyen and L. T. Pham, “Hybrid CNN-GRU model for physiological signal-based blood pressure estimation,” Biomedical Signal Processing and Control, vol. 71, p. 103062, 2022.
[12] K. J. Lee, S. H. Lim, and Y. J. Cho, “Transfer learning for health risk prediction: improving hypertension models with pre-trained representations,” IEEE Access, vol. 10, pp. 5120–5132, 2022.
[13] M. Y. Wang, Z. Q. Liu, and J. F. Huang, “Explainable GRU models for early-stage hypertension screening,” Journal of Medical Systems, vol. 46, no. 4, p. 77, 2022.
[14] P. S. Roy and A. Gupta, “Comparative analysis of LSTM and GRU for cardiovascular event prediction,” Expert Systems with Applications, vol. 184, p. 115550, 2022.
[15] C. H. Tan, J. L. Ong, and S. R. Lim, “Adaptive learning rates in Adam for clinical deep learning,” IEEE Trans. Neural Networks and Learning Systems, vol. 33, no. 5, pp. 2214–2226, 2022.
[16] F. R. Silva and T. O. Almeida, “Data augmentation and augmentation-aware training for tabular clinical datasets,” Information Sciences, vol. 587, pp. 125–140, 2022.
[17] D. S. Patel, R. N. Desai, and S. J. Mehta, “A compact GRU architecture for edge deployment of health prediction models,” IEEE Internet of Things Journal, vol. 9, no. 6, pp. 4620–4630, 2022.
[18] Y. K. Sun, P. H. Chen, and M. L. Yu, “Multimodal fusion of questionnaire and wearable data for blood pressure risk prediction,” Sensors, vol. 22, no. 1, p. 108, 2022.
[19] V. S. Kumar and A. R. Prasad, “Robust hypertension prediction using feature engineering and deep recurrent networks,” Applied Intelligence, vol. 52, no. 11, pp. 12230–12246, 2022.
[20] L. G. Martins and H. J. Sousa, “Cross-validation strategies for imbalanced clinical tasks: a case study on hypertension,” Knowledge-Based Systems, vol. 255, p. 109888, 2023.
[21] T. L. Nguyen, H. P. Tran, and G. H. Vu, “GRU with attention for personalized cardiovascular risk scoring,” IEEE J. Translational Engineering in Health and Medicine, vol. 11, p. e00077, 2023.
[22] S. R. Johnson and K. M. Lee, “Predictive modeling of hypertension onset using electronic health record sequences,” Journal of the American Medical Informatics Association, vol. 30, no. 3, pp. 601–611, 2023.
[23] M. I. Delgado and R. F. Ortiz, “End-to-end recurrent models for identifying hypertension phenotypes,” Computers in Biology and Medicine, vol. 154, p. 106506, 2023.
[24] H. P. Zhang, C. J. Wu, and J. X. Ma, “Lightweight GRU models for mobile health applications in low-resource settings,” IEEE Access, vol. 11, pp. 33010–33022, 2023.
[25] A. B. Santos and F. L. Ribeiro, “Feature importance and SHAP explanations for hypertension classifiers,” Machine Learning in Healthcare, vol. 5, pp. 45–59, 2023.
[26] R. M. Carter, P. Q. Lin, and J. E. Davis, “Evaluating optimizers for training health care neural networks: Adam, AdamW, and variants,” Neural Computing and Applications, vol. 35, pp. 1265–1281, 2023.
[27] K. T. Huang, L. Y. Zhao, and Y. M. Chen, “Temporal convolution and GRU hybrid model for blood pressure trend forecasting,” Sensors, vol. 23, no. 4, p. 1570, 2023.
[28] E. F. Oliveira, M. R. Sousa, and T. K. Lima, “Privacy-preserving federated GRU training for hypertension risk across clinics,” IEEE Trans. on Information Forensics and Security, vol. 18, pp. 1024–1036, 2023.
[29] S. A. Brown and L. J. Green, “Clinical validation of deep learning hypertension screening tools in community settings,” BMC Medical Informatics and Decision Making, vol. 23, p. 89, 2023.
[30] P. V. Rao, S. X. Zhang, and N. K. Mehta, “Regularization strategies for recurrent models in medical datasets,” Pattern Recognition, vol. 140, p. 109468, 2024.
[31] H. J. Karim, A. S. Ali, and L. P. Gomes, “Explainable AI for hypertension prognosis: combining SHAP and recurrent networks,” IEEE Access, vol. 12, pp. 7845–7858, 2024.
[32] Z. Q. Li and Y. F. Zhou, “A hybrid GRU-transformer approach for patient risk scoring,” Journal of Biomedical Informatics, vol. 140, p. 104313, 2024.
[33] M. N. Iqbal and S. T. Rahman, “Automated feature selection for clinical time series using mutual information and GRU,” Information Fusion, vol. 94, pp. 1–15, 2024.
[34] D. W. Kim and J. H. Park, “Impact of data quality and missingness on hypertension prediction models,” Journal of Medical Systems, vol. 48, no. 2, p. 21, 2024.
[35] Y. L. Cheng, H. B. Tan, and R. S. Lim, “Benchmarking recurrent architectures for blood pressure estimation,” IEEE Trans. Instrumentation and Measurement, vol. 73, pp. 1–12, 2024.
[36] F. M. Santos, P. D. Almeida, and R. G. Nunes, “Integrating wearables and survey data for continuous hypertension risk monitoring,” Sensors, vol. 24, no. 3, p. 678, 2024.
[37] L. A. Moreno and K. F. Silva, “Data balancing techniques impact on GRU model performance for hypertension,” Applied Soft Computing, vol. 142, p. 109107, 2024.
[38] J. P. Ortega, S. W. Lin, and T. H. Nguyen, “Fast-converging Adam variants for healthcare deep learning,” Neurocomputing, vol. 570, pp. 36–52, 2024.
[39] A. G. Pereira, V. H. Costa, and M. L. Ferreira, “Cross-population generalizability of deep hypertension predictors,” IEEE J. Biomed. Health Inform., vol. 28, no. 4, pp. 1890–1902, 2024.
[40] E. R. Santos and N. M. Oliveira, “Real-world deployment considerations for GRU-based clinical models,” JMIR Medical Informatics, vol. 12, no. 1, p. e42500, 2024.
[41] H. S. Yoon, D. K. Park, and H. G. Choi, “Lightweight recurrent models for wearable-based hypertension screening,” IEEE Internet of Things Journal, vol. 11, no. 8, pp. 6245–6257, 2024.
[42] R. L. Johnson and T. M. Wallace, “Multicenter evaluation of AI-based hypertension triage tools,” Lancet Digital Health, vol. 6, no. 1, pp. 34–44, 2024.
[43] S. V. Kumar, R. P. Sharma, and I. K. Nair, “Future directions in AI-driven cardiovascular risk prediction: challenges and opportunities,” IEEE Rev. Biomed. Eng., vol. 18, pp. 1–18, 2025.
