Prognostic Supervised Machine Learning Approaches for Upcoming Nigerian Asthma Incidence

Abstract

The utility of predictive models for the prognosis of asthma disease that rely on clinical history and findings has been on the constant rise, owing to the attempts to achieve better disease outcomes through improved clinical processes. With the prognostic model, the primary focus is on the search for a combination of features that are as robust as possible in predicting the disease outcome. Asthma remains one of the most prevalent chronic respiratory diseases worldwide, with increasing incidence in low- and middle-income countries such as Nigeria, where environmental factors play a substantial role in exacerbating disease outcomes. This study presents a hybrid machine learning framework for predicting asthma case patterns using environmental and meteorological data spanning January 2020 to December 2024. The proposed system integrates Facebook Prophet for time-series forecasting, Logistic Regression for binary risk classification, and an ensemble regression model combining Random Forest and Histogram Gradient Boosting for continuous prediction. A synthetically generated dataset reflecting Nigerian climatic conditions and asthma case counts was used to simulate real-world environmental dynamics. Feature engineering introduced interaction variables to capture complex dependencies among predictors, while data preprocessing ensured statistical integrity and model robustness. The Logistic Regression model achieved an accuracy of 70% and a ROC-AUC score of 0.875, demonstrating strong discriminatory capacity between high- and low-risk asthma periods. The ensemble regression model attained an R² of 0.72, RMSE of 15.6, and MAPE of 8.5%, indicating reliable predictive performance. Prophet’s forecast accurately captured seasonal trends consistent with Nigeria’s climatic cycles, identifying recurrent peaks during the rainy season months. The results highlight the effectiveness of combining time-series forecasting with supervised learning for environmental health prediction.  This research contributes to the development of AI-driven health intelligence systems capable of informing asthma control policies, environmental regulation, and preventive healthcare strategies across Nigeria and similar contexts.