Data mining and Machine learning based maternal risk prediction system

This study successfully demonstrated the application of machine learning for maternal health risk prediction. Through a comparative analysis of K-Nearest Neighbours, Naive Bayes, and Decision Tree classifiers, we established that the Decision Tree model provides the best performance, achieving an accuracy of 96.6%. Its ability to model complex relationships and provide interpretable results makes it an ideal candidate for a clinical decision support system. While promising, this work has limitations. The model was trained on a specific dataset and its generalizability should be tested on larger, more diverse populations. Future work should focus on: Exploring more advanced ensemble models like Random Forest and Gradient Boosting, which often build upon the strengths of Decision Trees. Incorporating a wider range of features, including lifestyle factors, socioeconomic data, and past medical history.   Developing and deploying a user-friendly application to make this predictive tool accessible to healthcare professionals in real- world settings.   Ultimately, machine learning stands as a powerful ally in the global effort to improve maternal health outcomes, offering data-driven insights to protect the well-being of mothers everywhere.

[1]     World Health Organization, "Maternal mortality”,22-February-2023. https://www.who.int/news-room/fact- sheets/detail/maternal-mortality [2]     A. S. Panicker and A. P. S. Kumar, "Maternal Risk Prediction Using Machine Learning and Deep Learning," in 2023 7th International Conference on Trends in Electronics and Informatics (ICOEI), 2023. Link: https://ieeexplore.ieee.org/abstract/document/1 0126027  [3]  A. Akbulut, A. E. Ertugrul, and V. Topcu, "A new maternal health risk dataset for machine learning and its effective utilization," Health Informatics Journal, vol. 29, no. 2, 2023. Link: https://journals.sagepub.com/doi/full/10.1177/ 14604582231172545 Approach for Maternal and Fetal Health Prediction," in 2023 5th International Conference on Inventive Research in Computing Applications (ICIRCA), 2023. Link: https://ieeexplore.ieee.org/abstract/document/1 0221356 [9] S. Ahmed, A. E. E. Emon, S. T. R. Shah, and V. Topcu, "Maternal Health Risk Data Set," UCI Machine Learning Repository, 2020. Link: https://archive.ics.uci.edu/dataset/863/maternal +health+risk+data+set [4]   A. E. Johnson, T. J. Pollard, L. Shen, H. Li-wei, M. Feng, and R. G. Mark, "MIMIC- III, a freely accessible critical care database," Scientific Data, vol. 3, 2016. (Often used for such studies, this paper describes the database). Link: https://www.nature.com/articles/sdata201635 [5]  M. R. Devi, S. K. B. S. S, N. R. M, and P. N, "Maternal Health-Risk Prediction using Decision Tree and SVM Algorithms," in 2022 International Conference on Advancements in Electrical, Electronics,      Communication, Computing and Automation (ICAECA), 2022. Link: https://ieeexplore.ieee.org/abstract/document/9 952219 [10] H. Koivu, S. Sairanen, M. Gissler, P. Stefanovic, and P. K. Heinonen, "Machine learning reveals complex risk factor patterns for severe maternal morbidity," Acta Obstetricia et Gynecologica Scandinavica, vol. 101, no. 6, pp. 691-699, 2022. Link: https://obgyn.onlinelibrary.wiley.com/doi/full/ 10.1111/aogs.14364