A Diabetes Prediction Decision Support System Using Machine Learning

Abstract

Diabetes is a chronic metabolic disorder affecting millions worldwide and is a leading cause of serious health complications such as heart disease, kidney failure, and vision loss. Early detection and accurate risk prediction are vital for effective disease management and reduction of long-term health burdens. This study presents a robust diabetes prediction decision support system leveraging advanced machine learning techniques applied to a curated dataset obtained from Kaggle. The primary objective is to address the critical need for early and accurate diabetes diagnosis by developing a reliable and efficient prediction model. Through comprehensive data preprocessing, detailed feature engineering, and systematic model training, a Random Forest Classifier was developed and optimized using GridSearchCV. The resulting model achieved an impressive accuracy of 94.70%, demonstrating a precision of 0.98, recall of 0.96, and F1-score of 0.97 for non-diabetic cases (class 0), and a precision of 0.65, recall of 0.80, and F1-score of 0.72 for diabetic cases (class 1). These results underscore the model's high reliability in detecting non-diabetic instances and its reasonable effectiveness in identifying diabetic cases. The findings offer significant implications for clinical decision-making, enabling healthcare professionals to implement proactive interventions and personalized treatment strategies. Furthermore, the model provides a valuable foundation for future enhancements in medical data analysis and health informatics applications aimed at improving chronic disease management.

References

[1] H. Wu, S. Yang, Z. Huang, J. He, and X. Wang, “Type 2 diabetes mellitus prediction model based on data mining,” Inform. Med. Unlocked, vol. 10, pp. 100–107, 2018.

[2] M. M. Islam, R. Ferdousi, S. Rahman, and H. Y. Bushra, “Likelihood prediction of diabetes at early stage using data mining techniques,” in Computer Vision and Machine Intelligence in Medical Image Analysis. Singapore: Springer, 2020, pp. 113–125.

[3] F. G. Woldemichael and S. Menaria, “Prediction of diabetes using data mining techniques,” in 2018 2nd Int. Conf. Trends Electron. Inform. (ICOEI), May 2018, pp. 414–418.

[4] C. Fiarni, E. M. Sipayung, and S. Maemunah, “Analysis and prediction of diabetes complication disease using data mining algorithm,” Procedia Comput. Sci., vol. 161, pp. 449–457, 2019.

[5] A. Aldallal and A. A. A. Al-Moosa, “Using data mining techniques to predict diabetes and heart diseases,” in 2018 4th Int. Conf. Frontiers Signal Process. (ICFSP), Sep. 2018, pp. 150–154.

[6] F. A. Khan, K. Zeb, M. Al-Rakhami, A. Derhab, and S. A. C. Bukhari, “Detection and prediction of diabetes using data mining: a comprehensive review,” IEEE Access, vol. 9, pp. 43 711–43 735, 2021.

[7] I. Kavakiotis, O. Tsave, A. Salifoglou, N. Maglaveras, I. Vlahavas, and I. Chouvarda, “Machine learning and data mining methods in diabetes research,” Comput. Struct. Biotechnol. J., vol. 15, pp. 104–116, 2017.

[8] P. Kumar, P. Kumar, A. Srivastava, V. D. A. Kumar, K. Vengatesan, and A. Singhal, “Comparative analysis of data mining techniques to predict heart disease for diabetic patients,” in Adv. Comput. Data Sci., Singapore: Springer, Apr. 2020, pp. 507–518.

[9] T. R. Mahesh et al., “Blended ensemble learning prediction model for strengthening diagnosis and treatment of chronic diabetes disease,” Comput. Intell. Neurosci., vol. 2022, 2022.

[10] N. Oza and A. Bokhare, “Diabetes prediction using logistic regression and K-nearest neighbor,” in Congr. Intell. Syst., Singapore: Springer, 2022, pp. 407–418.

[11] K. S. Anil and R. Jain, “Data mining techniques in diabetes prediction and diagnosis: a review,” in 2022 6th Int. Conf. Trends Electron. Inform. (ICOEI), Apr. 2022, pp. 1696–1701.