This project aims to develop novel statistical modeling approaches and parametric/nonparametric inference methods for the analysis of high-dimensional spatio-temporal data. Some parts of the project are directly motivated by some applications, particularly in the fields of astronomy, food safety, and wastewater treatment. The project is divised into four core axes : the first part focuses on the segmentation and classification of satellite remote sensing data to detect changes such as vegetation type shifts or wildfires. The second part investigates dynamical multi-compartment models to study the relation between the external exposure contamination and the bio-distribution of contaminants in different parts of the body. The third part proposes new non-parametric estimators in both non-parametric and linear regression functional models. The fourth part deals with spatio-temporal models using Gaussian process prior to infer the presence of drugs and pathogens in sewage systems and wastewater treatment plants. This project is coordinated by Karine Bertin (Universidad de Valparaíso, Chile), Aldo Medina (Federal University of Pernambuco, Brasil) and Emilie Lebarbier (Université Paris Nanterre, Francia).
