Bogotá - Ciencias - Maestría en Ciencias - Estadística · 2023
Classification techniques for imaginary speech brain signal through spatial functional data
The present work aims to classify the thought of the five Spanish vowels measured by electroencephalograms (EEG) of 21 electrodes around the Broca's area of the brain of 23 individuals. This was addressed by the framework of spatial functional data, considering each EEG a continuous curve in L2 and performing functional kriging, several images were constructed to apply classification techniques of machine and deep learning. Finally, both classification routines on average achieve more than 91\% precision for each individual, considering that each individual should have its own classification mechanism. (Texto tomado de la fuente)