Improving Public Health Policies with Indoor Air Quality Predictive Models

Indoor air quality is important for public health. This study was designed to develop predictive models focusing on indoor air quality, specifically targeting levels of CO2, TVOC, PM2.5, and PM10. We implemented and trained Machine Learning Models-Regression Forest Model and Gradient-Boosted Tree Model-using a dataset from the states of Puebla and Morelos in Mexico. The dataset incorporated various environmental variables, including pollutant levels, temperature, relative humidity, population density, and ventilation characteristics, all of which were found to significantly influence the presence of indoor air contaminants. These findings are instrumental in formulating policies to mitigate poor indoor air quality. Moreover, the study suggests that it is feasible to predict when contaminants will reach harmful levels by monitoring changes in these variables.