Predicting Ozone Layer Concentration Using Multivariate Adaptive Regression Splines, Random Forest and Classification and Regression Tree

Abstract

Air pollution is one of the major environmental worries in recent time. Abrupt increase in the concentration of any gas leads to air pollution. The cities are mostly affected due to the abundance of population there. One of the worst gaseous pollutants is OZONE (O3). In this paper, we propose three predictive models for estimation of concentration of ozone gases in the air which are Random Forest, Multivariate Adaptive Regression Splines and Classification and Regression Tree. Evaluation of the prediction models indicates that the Multivariate Adaptive Regression Splines model describes the dataset better and has achieved significantly better prediction accuracy as compared to the Random Forest and Classification and Regression Tree. A detailed comparative study has been carried out on the performances of Random Forest, Multivariate Adaptive Regression Splines and Classification and Regression Tree. MARS gives the result by considering less variables as compared to other two. Moreover, Random Forest takes a little more time for building the tree as the elapsed time was calculated to 45 s in this case. In addition, variable importance for each model has been predicted. Observing all the graphs Multivariate Adaptive Regression Splines gives the closest curve of both train and test set when compared. It can be concluded that multivariate adaptive regression splines can be a valuable tool in predicting ozone for future.