A constrained grouped additive index model, applied for weather-health warning systems

In the last twenty years, a number of countries implemented heat-health warning system to monitor heat waves and take action to reduce their impact on the population.

Such warning systems often rely on two components, indices representing the monitored environmental phenomenon of interest and thresholds above which the phenomenon is considered to induce a high risk on human health. The work presented here focuses on the former (indices). In this regard, we introduce a novel and general method to construct one or several indices.

In order to construct interpretable indices with good prediction performances, the proposed model uses several constraints and a grouping of the exploratory variables. Hence the model is called constrained grouped additive index model (c-GAIM). We show the use of c-GAIM to construct indices for the heat-health warning system of the province of Quebec.

Please register for the event.

About Fateh Chebana

Fateh Chebana holds a PhD in Mathematical statistics from Université Paris 6, France (2003). Since 2010, he is professor at INRS (Institut national de la recherche scientifique) within a multidisciplinary research centre Water-Earth and Environment. His main recent research work focuses on statistical hydrology (the modeling and estimation of hydrological events in a multivariate context), as well as on environmental health methods and applications.