Beyond One Room: Comprehensive Predictive Analysis of CO2 in Indoor Air Quality

Indoor air quality is important for public health. This study was designed to develop predictive models that concentrate on indoor air quality, focusing on the CO2 concentrations. Implementing and training the Machine Learning Models—Regression Forest Model and Gradient-Boosted Tree Model—on a dataset of measurements in Mexico and other sources having pollutant levels, temperature, relative humidity, people density, and ventilation characteristics. The models used had many scenarios of relative humidity, temperature and pollutant levels, demonstrating the relation between the characteristics of the space, human activity and indoor CO2 concentration. The result was a model with acceptable accuracy, predicting CO2 levels in indoor spaces.