OPTIMIZING CONVOLUTIONAL NEURAL NETWORKS WITH SIMULATED ANNEALING FOR HEART DISEASE PREDICTION

Year 2025, Volume: 13 Issue: 4, 1023 - 1033, 30.12.2025

Osama Burak Elhalid , Mehmet Fatih Demiral

https://doi.org/10.21923/jesd.1638469

Abstract

Cardiovascular diseases (CVDs) remain the leading cause of death worldwide, underscoring the urgent need for reliable predictive models that can support early diagnosis and effective treatment. This study introduces a novel framework that combines Convolutional Neural Networks (CNNs) with the Simulated Annealing (SA) algorithm to optimize critical hyperparameters, including the number of filters, kernel size, hidden units, and batch size. The experiments were conducted on the publicly available Cleveland Heart Disease dataset from the UCI Machine Learning Repository, which contains 303 patient records with 14 clinical attributes. The proposed SA-CNN model achieved an accuracy of 96.1% and an F1-score of 0.96, surpassing baseline CNNs and traditional optimization techniques such as grid search and random search. By systematically navigating the hyperparameter space, the SA algorithm reduced overfitting and improved the model’s generalization ability. These findings highlight the effectiveness of metaheuristic optimization in enhancing deep learning models for medical diagnosis and provide a robust, scalable framework for AI-driven heart disease prediction.

Keywords

Simulated Annealing , Convolutional Neural Networks , Heart Disease Prediction

KALP HASTALIĞI TAHMİNİ İÇİN SİMÜLE EDİLMİŞ TAVLAMA İLE EVRİŞİMSEL SİNİR AĞLARININ OPTİMİZE EDİLMESİ

Abstract

Kardiyovasküler hastalıklar (KVH'ler) dünya çapında önde gelen ölüm nedeni olmaya devam etmekte ve erken tanı ve etkili tedaviyi destekleyebilecek güvenilir tahmin modellerine acil ihtiyaç olduğunu vurgulamaktadır. Bu çalışma, filtre sayısı, çekirdek boyutu, gizli birimler ve parti boyutu dahil olmak üzere kritik hiperparametreleri optimize etmek için Evrişimli Sinir Ağları'nı (CNN'ler) Simüle Edilmiş Tavlama (SA) algoritmasıyla birleştiren yeni bir çerçeve sunmaktadır. Deneyler, 14 klinik özelliğe sahip 303 hasta kaydı içeren UCI Makine Öğrenmesi Deposu'ndan halka açık Cleveland Kalp Hastalığı veri kümesi üzerinde yürütülmüştür. Önerilen SA-CNN modeli, %96,1'lik bir doğruluk ve 0,96'lık bir F1 puanı elde ederek, temel CNN'leri ve ızgara araması ve rastgele arama gibi geleneksel optimizasyon tekniklerini geride bırakmıştır. SA algoritması, hiperparametre alanında sistematik olarak gezinerek aşırı uyumu azaltmış ve modelin genelleme yeteneğini geliştirmiştir. Bu bulgular, tıbbi teşhis için derin öğrenme modellerini geliştirmede metasezgisel optimizasyonun etkinliğini vurgulamakta ve yapay zeka destekli kalp hastalığı tahmini için sağlam, ölçeklenebilir bir çerçeve sağlamaktadır.

Keywords

Simüle Edilmiş Tavlama , Konvolüsyonlu Sinirsel Ağlar , Kalp Hastalığı Tahmini

References

• Aarts, E. H. L., & Van Laarhoven, P. J. M. (1989). Simulated annealing: an introduction. Statistica Neerlandica, 43(1), 31-52.
• Alzubaidi, L., Al-Shamma, O., & Fadhel, M. A. (2023). Genetic Algorithm-based Convolutional Neural Network Feature Engineering for Coronary Heart Disease Prediction. Journal of Biomedical Informatics, 135, 104195. https://doi.org/10.1016/j.jbi.2023.104195
• Bergstra, J., & Bengio, Y. (2012). Random search for hyper-parameter optimization. Journal of Machine Learning Research, 13(10), 281–305.
• Chopard, B., Tomassini, M., Chopard, B., & Tomassini, M. (2018). Simulated annealing. An introduction to metaheuristics for optimization, 59-79.
• Elhalid, O. B., & Isık, A. H. (2024). ENHANCING MEDICAL OFFICER SCHEDULING IN HEALTHCARE ORGANIZATIONS: A COMPREHENSIVE INVESTIGATION OF GENETIC AND GOOGLE OR TOOLS ALGORITHMS FOR MULTI-PROJECT RESOURCE-CONSTRAINED OPTIMIZATION. International Journal of 3D Printing Technologies and Digital Industry, 8(1), 92-103. https://doi.org/10.46519/ij3dptdi.1415512
• Elhalid, Osama Burak, and Alm Alhelal, Zaynelabdin, and HASSAN, SAMER, Exploring the Fundamentals of Python Programming: A Comprehensive Guide for Beginners (October 25, 2023). Available at SSRN: https://ssrn.com/abstract=4612765 or http://dx.doi.org/10.2139/ssrn.4612765
• Eskicioğlu, Ö. C., Dolićanin, E., Işık, A. H., & Rifai, K. (2021). Recognition and detection with deep learning methods. Scientific Publications of the State University of Novi Pazar Series A: Applied Mathematics, Informatics and Mechanics, 13(2), 105-115.
• Greening, D. R. (1990). Parallel simulated annealing techniques. Physica D: Nonlinear Phenomena, 42(1-3), 293-306.
• Gu, J., Wang, Z., Kuen, J., Ma, L., Shahroudy, A., Shuai, B., ... & Chen, T. (2018). Recent advances in convolutional neural networks. Pattern Recognition, 77, 354–377. https://doi.org/10.1016/j.patcog.2017.10.013
• Guo, Z., & Cao, Y. (2022, October). SA-CNN: Application to text categorization issues using simulated annealing-based convolutional neural network optimization. In Proceedings of the 2022 6th International Conference on Electronic Information Technology and Computer Engineering (pp. 1932-1939).
• Gülcü, A., & Kuş, Z. (2021). Multi-objective simulated annealing for hyper-parameter optimization in convolutional neural networks. PeerJ Computer Science, 7, e338.
• Işik, A. H., Ersoy, M., Köse, U., Türkçetin, A. Ö., & Çolak, R. (2021). Deep Learning Based Classification Method for Sectional MR Brain Medical Image Data. In Trends in Data Engineering Methods for Intelligent Systems: Proceedings of the International Conference on Artificial Intelligence and Applied Mathematics in Engineering (ICAIAME 2020) (pp. 669-679). Springer International Publishing.
• Janosi, A., Steinbrunn, W., Pfisterer, M., & Detrano, R. (1989). Heart Disease [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C52P4X
• Kirkpatrick, S., Gelatt, C. D., & Vecchi, M. P. (1983). Optimization by simulated annealing. Science, 220(4598), 671–680. https://doi.org/10.1126/science.220.4598.671
• Kora, P., & Kalva, S. K. (2024). Enhanced heart disease prediction through hybrid CNN-TLBO-GA optimization. Cogent Engineering, 11(1), 2384657. https://doi.org/10.1080/23311916.2024.2384657
• Kuo, C. L., Kuruoglu, E. E., & Chan, W. K. V. (2022). Neural network structure optimization by simulated annealing. Entropy, 24(3), 348.
• LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
• Ogunpola, A., Saeed, F., Basurra, S., Albarrak, A. M., & Qasem, S. N. (2024). Machine learning-based predictive models for the detection of cardiovascular diseases. Diagnostics, 14(2), 144.
• Polepaka, S., Ram Kumar, R. P., Palakurthy, D., Manasa, V., Saritha, A., Dixit, S., ... & Adnan, M. M. (2024). Optimized a convolutional neural network using the grasshopper optimization technique for enhanced heart disease prediction. Cogent Engineering, 11(1), 2423847.
• Rere, L. R., Fanany, M. I., & Arymurthy, A. M. (2015). Simulated annealing algorithm for deep learning. Procedia Computer Science, 72, 137-144.
• Şengöz, N., Yiğit, T., Özmen, Ö., & Isık, A. H. (2022). Importance of preprocessing in histopathology image classification using a deep convolutional neural network. Advances in Artificial Intelligence Research, 2(1), 1-6.
• Tan, C. M. (Ed.). (2008). Simulated annealing. BoD–Books on Demand.
• Van Laarhoven, P. J., & Aarts, E. H. (1987). Simulated annealing: Theory and applications. Springer.
• World Health Organization (WHO). (2023). Cardiovascular diseases (CVDs). Retrieved from https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(vcds)