Abstract:: The main goal of this proposal is to develop statistical/machine learning methodologies for the analysis of data with complex structures, based on the applied experience acquired in the development of the FONDEF 1141057 project. This research focuses on six aspects: (I) flexible mixed models for modeling trajectories in complex domains; (II) develop flexible cure rate models using machine learning algorithms such as random forests or neural networks; (III) developing spatial survival models using copulas in order to capture flexible dependences; (IV) develop efficient deep learning algorithms using multimodal data; (V) extract important information from patient interventions or writings using natural language processing techniques in Spanish; and (VI) applications of the proposed models in some interesting and motivating scenarios.
Director: Rolando de la Cruz
Co-Director: Claudia Duran-Aniotz
PIs: Agustín Ibañez, Gonzalo Ruz y Moreno Bevilacqua
Support: ANID/Anillo
