Aprendizaje Automática para la Predicción del Desempeño de la Memoria de Trabajo

General objective Analyze machine learning methodologies through the implementation of algorithms that contribute to the characterization and prediction of WM performance. Justification Nowadays, the importance of data analysis in any human activity is evident. From medicine to entertainment, it generates data that needs to be analyzed. In the field that concerns us, that of EEG signals, data analysis plays an important role. That is why this project seeks to develop and implement efficient and agile methodologies for managing variability in EEG signals and their application in WM. Our goal is to implement and explore machine learning methodologies that contribute to the prediction and characterization of performance. of working memory. Methodologies that allow us to manage the intrinsic variability of EEG signals and propose solutions that help usability and improve performance prediction; problems that are still a challenge in our field. In a first phase of this project, it will allow us to know and understand the methods that are currently used to control the variability of the data. In a second stage we plan to develop new machine learning algorithms focused on the aforementioned problem. This project aims to analyze a common problem in machine learning so that the solutions achieved can be extrapolated to other research areas. Likewise, this project will contribute to promoting “data science” research at UPS through the innovation and implementation of feature extraction and machine learning techniques.