Machine Learning
Publications
- M. Carrasco, B. Ivorra, J. López, A. M. Ramos. Embedded feature selection for robust probability learning machines. Pattern Recognition. 159. 2025. DOI: 10.1016/j.patcog.2024.111157
Preprints UCM
- Robust SVM Classification with lp-Quasi-Norm Feature Selection
M. Carrasco, J. López, B.P.P. Ivorra, M. Marechal, Á.M. Ramos del Olmo
https://docta.ucm.es/entities/publication/01f67cea-6743-4514-b61d-a9d75835c114 - Robust SVM Classification with l0-Norm Feature Selection
M. Carrasco, B.P.P. Ivorra, J. López, Á.M. Ramos del Olmo
https://docta.ucm.es/entities/publication/19b748ef-6284-4078-adb7-67a91ad1fc65 - A Note on Probability and Conditional Probability SVM Models
M. Carrasco, J. López, B.P.P. Ivorra, Á.M. Ramos del Olmo
https://docta.ucm.es/entities/publication/32f005eb-4f5a-4645-b546-8e1d5aefe2a6
News
- March 6 2025. Unidad de Cultura Científica. Oficina de Transferencia de Resultados de Investigación (OTRI). A research team led by the IMI researchers of the MOMAT Group Benjamin Ivorra and Ángel Ramos, in collaboration with colleagues from the Chilean Universities of Los Andes and Diego Portales, has developed a novel approach using advanced mathematical techniques to enhance the accuracy and efficiency of AI models. Their work focuses on refining data selection strategies to improve the reliability of case classification. By optimizing data inputs, the AI can more effectively analyze new information and make precise classifications—for instance, assessing whether a patient faces a high risk of cancer or identifying potentially fraudulent banking transactions. Published in Pattern Recognition, the study demonstrates how removing irrelevant data maintains prediction quality while making AI models faster, more interpretable, and more resilient to variations in input data. See also https://www.ucm.es/otri/noticias-mejora-ia-aplicaciones-cancer-ucm.