Modelo de clasificación de señales EEG usando Random Forest

Abstract

Emotions are one of the components of human beings that are essentials in daily activities, like interaction with other people, decision making and learning. Due to what was exposed, it is important to detect, recognize and interpret emotions using computer systems to improve communication between people and machines, which facilitates to computers to understand communication between humans. This study proposes the generation of a model which allows the classification of people's feelings, from their EEG signals. To accomplish this objective, the “brain-computer interface EMOTIV EPOC” was used, which allowed the collection of electroencephalographic information from 50 persons. The obtained information was stored in a database for the generation of a model with its respective classification analysis. To predict three moods (happy, sad and relaxed) of the signals of any person, a model was generated using Random Forest, which got accuracies of 97.21% for joy, 76% for relaxation and 76% for sadness. Finally, the model was used to generate a real-time prediction algorithm for the three emotions, which captures the person's EEG signals, then executes the generated algorithm and displays the result on the screen through representative images of each emotion.