BIOSENSADO PARA LA DETECCIÓN DE EMOCIONES; CLASIFICACIÓN DE EVENTOS PARA DOS TECNOLOGÍAS.

ALVÍDREZ LOZANO, FLORENTINO (2019) BIOSENSADO PARA LA DETECCIÓN DE EMOCIONES; CLASIFICACIÓN DE EVENTOS PARA DOS TECNOLOGÍAS. Maestría thesis, Universidad Autónoma de Chihuahua.

[img]
Vista Previa
Text
BIOSENSADO PARA LA DETECCIÓN DE EMOCIONES - CLASIFICACIÓN DE EVENTOS PARA DOS TECNOLOGÍAS _ FLORENTINO ALVÍDREZ.pdf

Download (4045Kb) | Vista Previa

Resumen

Emotions are a fundamental part of our way of thinking and our behavior. They themselves trigger physiological states that can be measured and classified in order to identify the type of emotion a person experiences. In this work, several pictures from the IAPS (International Affective Picture System) data set were used to evoke different types of emotions to 18 participants divided into two groups (9 for each one). Two types of sensing technologies (e-Health and Bitalino) were used in the study with the aim of using physiological signals such as heart rate and electrodermal activity to identify and classify high and low Valence, as well high and low Arousal. The classifiers implemented were Random-Forest, Support vector machine (SVM), Neural Network (MLP), K-Nearest Neighbors (KNN) and Decision-Tree. As a result it were possible to identify with F1-Score an accuracy of 51.06% (with Decision-Tree) in Bitalino and 54.81% (with KNN) in e-Health for high and low Valence, while to detect Arousal, it was achieved 70.67% (with Random-Forest) in Bitalino and 55.85% (with KNN) in e-Health.

Tipo de Documento: Tesis (Maestría)
Palabras Clave: Bitalino, E-Health, Biosensors, IAPS, Machine Learning
Clasificación temática: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Usuario Remitente: Admin Administrador del Respositorio
Depositado: 19 Nov 2019 18:48
Ultima Modificación: 19 Nov 2019 18:48
URI: http://repositorio.uach.mx/id/eprint/247

Actions (login required)

Ver Objeto Ver Objeto

Universidad Autonoma de Chihuahua