Comparison of Different Regularized and Shrinkage Regression Methods to Predict Daily Tropospheric Ozone Concentration in the Grand Casablanca Area

International audience Tropospheric ozone (O3) is one of the pollutants that have a significant im-pact on human health. It can increase the rate of asthma crises, cause perma-nent lung infections and death. Predicting its concentration levels is thereforeimportant for planning atmospheric protection strategies. The aim of thisstudy is to predict the daily mean O3 concentration one day ahead in theGrand Casablanca area of Morocco using primary pollutants and meteoro-logical variables. Since the available explanatory variables are multicollinear,multiple linear regressions are likely to lead to unstable models. To counte-ract the multicollinearity problem, we compared several alternative regressionmethods: 1) Continuum Regression; 2) Ridge & Lasso Regressions; 3) Prin-cipal component regression (PCR); 4) Partial least Square regression & sparsePLS and; 5) Biased Power Regression. The aim is to set up a good predictionmodel of the daily ozone in the Grand Casablanca area. These models are fit-ted on a training data set (from the years 2013 and 2014), tested on a data set(from 2015) and validated on yet another data set data (from 2015). The Las-so model showed a better performance for the prediction of ozone concentra-tions compared to multiple linear regression and its other alternative me-thods.