Predicting Respiratory Diseases Attributed to PM2.5 Air Pollution in Nairobi County Using Random Forest Model

This study investigates the predictive capability of a Random Forest model in identifying respiratory diseases attributed to PM2.5 exposure in Nairobi County. Leveraging a comprehensive dataset encompassing demographic and air quality variables, the model demonstrated robust performance metrics, achieving an accuracy of 79.97% and an area under the curve (AUC) of 0.872. These results highlight the model's effectiveness in distinguishing between respiratory and cardiovascular conditions. The model's sensitivity and specificity were 81.88% and 73.27%, respectively, indicating a strong ability to correctly identify both true positives and true negatives. Analysis of feature importance revealed that age and PM2.5 concentrations were the most influential factors in predicting health outcomes, emphasizing the significant impact of air pollution and demographic factors on respiratory and cardiovascular health. Furthermore, the consistent train and test error rates across varying training set sizes suggest the model's stability and generalizability. This study underscores the importance of addressing air quality issues to mitigate the health impacts of PM2.5 exposure in urban settings.