Análisis de señales EEG para la detección de emociones, usando algoritmos de aprendizaje automático: reconocimiento de emociones usando características en los dominios del tiempo, frecuencia y tiempo-frecuencia extraídas de un dataset de EEG usando técnicas de aprendizaje automático.

Abstract

This document describes in detail the process of feature extraction used for electroencephalogram signals analysis and study using machine learning techniques and algorithms. From signal processing and their mathematical analysis in different domains such as: Time domain, Frequency domain and Time-Frequency domain, represented by a well-organized dataset over columns and features; through the reduction of these features using different techniques such as PCA Algorithm or the Correlation Matrix analysis; until the definition and comparison of distinct machine learning models that study those features. All the previous process allows finding new intelligent classification models for human emotion recognition, which have a great and spread importance in many fields related with mankind study and its inside.