On sparse ensemble methods: An application to short-term predictions of the evolution of COVID-19
In a recently published paper, Professor Dolores Romero Morales, Professor in Operations Research at Department of Economics, together with co-authors, developed a novel sparse forecasting ensemble method. Their methodology, combining mathematical optimization and machine learning, penalizes through a regularization term poor individual performance of some of the base regressors. With this they can successfully tackle dynamic environments where the composition of the ensemble as well as the weights given to the base regressors is bound to change to achieve competitive accuracies. They illustrate this approach with real-world data sets arising in the COVID-19 context in Denmark and Spain.
This paper is in open access and can be found following the link below.
This project has received funding from the European Union's Horizon 2020 research and Innovation programme under the Marie Skłodowska-Curie grant agreement No. 822214: H2020 MSCA RISE Action Network of European Data Scientists, www.riseneeds.eu.
By S. Benítez Peña, E. Carrizosa, V. Guerrero, M.D. Jiménez-Gamero, B. Martín-Barragán, C. Molero-Río, P. Ramírez-Cobo, D. Romero Morales, M.R. Sillero-Denamiel