Research Article
A Machine Learning Model for Resurgence Prediction of Ten Infectious Diseases in Senegal
@INPROCEEDINGS{10.1007/978-3-032-15154-4_6, author={Cheikh Tidiane Seck and Abdourahmane Ndao}, title={A Machine Learning Model for Resurgence Prediction of Ten Infectious Diseases in Senegal}, proceedings={Innovations and Interdisciplinary Solutions for Underserved Areas. 8th International Conference, InterSol 2025, Ouagadougou, Burkina Faso, July 3--4, 2025, Proceedings}, proceedings_a={INTERSOL}, year={2026}, month={11}, keywords={Resurgence prediction Multi-output decision tree Inertia criterion}, doi={10.1007/978-3-032-15154-4_6} }- Cheikh Tidiane Seck
Abdourahmane Ndao
Year: 2026
A Machine Learning Model for Resurgence Prediction of Ten Infectious Diseases in Senegal
INTERSOL
Springer
DOI: 10.1007/978-3-032-15154-4_6
Abstract
Ongoing changes in our environment have favored the emergence and/or resurgence of numerous infectious diseases, posing a real public health problem. Our aim in this work is to predict the resurgence of such diseases in the Senegalese context, using an extract from the Ministry of Health’s epidemiological surveillance database, comprising 68,698 observations. We propose a multi-output decision tree (MO-DT) model which, introduces an inertia criterion (calculated with the chi-squared distance) as the node impurity measure, and allows to simultaneously predict ten infectious diseases targeted by the surveillance program. The results show that these diseases have an average resurgence probability of 12.2%, with the exception of Poliomyelitis, which records a resurgence probability of 2.4%. Our study also reveals that during the period under consideration (January 2018 to November 2022), Covid-19 had a fairly high resurgence probability approaching 60%. In comparison with multi-class random forests (MC-RF) and multinomial logistic regression (MLR), we find that our model performs slightly better. For example, for Accuracy, we have: MO-DT (0.9945), MC-RF (0.9943), RLM (0.9162).