Examining the Impact of Environmental Factors on Asthma Development Using a Boosting Ensemble Model

Abstract

Asthma is a chronic inflammatory airway disease which is affected by both environmental, demographic, lifestyle and clinical factors. Early prediction of asthma risk is vital for early intervention and better asthma management. . This study suggests the boosting ensemble approach to predict the development of asthma on a dataset of 2,392 people with demographic information, lifestyle information, allergy history, environmental exposures, and clinical attributes. To enhance the reliability of the model, data preprocessing techniques such as data cleaning, encoding categorical features, scaling features, and feature selection were employed. The proposed framework uses a soft-voting approach that averages class probabilities and compares the models using a consistent experimental setup. The experimental results show that the proposed boosting ensemble produced an accuracy of 99%, better than all baseline models and provides highly accurate predictions for asthma and non-asthma classes. The results show that combining some boosting algorithms improves the prediction accuracy and robustness than using the individual classifiers. This proposed framework offers a reliable tool for predicting asthma in early life and may serve as a basis for future clinical decision making, but needs to be validated with other independent and geographically distinct populations in order to establish its generalizability. Because the dataset is synthetic, these results show methodological performance under controlled experimental conditions only; clinical usefulness would require validation on real patient data.