Sistema de detección de estados emocionales de usuarios según canciones escuchadas
Cargando...
Archivos
Fecha
2016
Autores
Barraza Mercado, Gustavo A.
Título de la revista
ISSN de la revista
Título del volumen
Editor
Ediciones Universidad Simón Bolívar
Facultad de ingenierías
Facultad de ingenierías
Resumen
La música, se ha convertido en un elemento básico en el diario vivir del ser humano, no solo como uno de los mecanismos regulador de las emociones, sino también como el medio por el cual se pueden expresar.
Actualmente, se ha dada gran atención a la detección de los estados emocionales en los seres humanos a través del análisis de parámetros fisiológicos y neurológicos, es así, como continuamente se busca desarrollar aplicaciones informáticas automatizadas que reciban información directa a través de la conexión humano-máquina, que cuente con entradas objetivas y reduzca las anomalías que se presentan en los métodos tradicionales, por el alto grado de libertad inherente a las respuestas suministradas por los usuarios. Es por esto que, existe un interés en el desarrollo de sistemas que permitan conocer el estado emocional de un usuario a partir de las canciones que escucha, sin que éste se someta al estrés de realizar un test o algún examen fisiológico que afecte el resultado de su estado emocional. Por lo tanto, en este proyecto de investigación, se abordó la detección y clasificación de los estados emocionales de un usuario mediante el uso de técnicas de aprendizaje supervisado basados en inteligencia artificial, las cuales permiten suplir las dificultades expuestas.
Los resultados que obtuvimos nos permiten emitir un juicio más acertado sobre el estado emocional que presenta el usuario, en el tiempo que utiliza el sistema, sea corto o largo y con una interacción sencilla y agradable. El principal objetivo fue el desarrollo de una metodología, que basado en un sistema reproductor de audio o multimedia, detecte el estado emocional del propietario o en su defecto de quien lo utilice, según las canciones seleccionadas y escuchadas; por eso se ha tomado como tema principal el uso de la música como medio de detección del estado emocional, y con ello dar a conocer la importancia de esta investigación en el aporte de futuros avances en la detección de los estados emocionales.
El sistema diseñado se validó mediante un conjunto de pruebas realizadas sobre escenarios de simulación y casos reales, las cuales manifestaron un excelente desempeño del sistema, dado que en todas las pruebas se detectó de forma correcta el estado emocional a partir de la lista de reproducción de las canciones escuchadas. Otro aporte destacable de este trabajo de investigación se basa en la creación de un (Dataset) para la detección de estados emocionales a partir de la música, el cual podría ser utilizado para determinar desordenes o trastornos emocionales en usuarios.
Music has become a staple in the daily life of the human being, not only as one of the mechanisms regulating emotions, but also as the means by which they can express. Currently, great attention has been given to the detection of emotional states in humans through the analysis of physiological parameters and neurological, so as continually seeks to develop automated applications that receive direct information through the humanmachine connection, objective that has inputs and reduce the anomalies that occur in traditional methods, the high degree of inherent to the answers provided by user's freedom. Thus, there is interest in the development of systems to meet the emotional state of a user from the songs they listen, without it is subjected to the stress of performing a test or a physiological test that affects the outcome of your emotional state. Therefore, in this research project, the detection and classification of the emotional states of a user by using supervised learning techniques based on artificial intelligence was addressed, which allow meet the above difficulties. The results we obtained allow us to make a more accurate about the emotional state that presents the user, in time using the system, either short or long, with a simple and pleasant interaction trial. The main objective was the development of a methodology, based on an audio or multimedia system, detect the emotional state of the owner or failing of who uses it, as selected and played songs; why it has taken as its main theme the use of music as a means of detecting the emotional state, and thereby raise awareness of the importance of this research in providing future advances in detecting emotional states. The designed system was validated through a set of tests on simulation scenarios and case studies, which demonstrated excellent performance of the system, since in all tests detected correctly the emotional state from the playlist the songs heard. Another important contribution of this research is based on the creation of a (Dataset) for detecting emotional states from the music, which could be used to determine disorders or emotional disorders in users.
Music has become a staple in the daily life of the human being, not only as one of the mechanisms regulating emotions, but also as the means by which they can express. Currently, great attention has been given to the detection of emotional states in humans through the analysis of physiological parameters and neurological, so as continually seeks to develop automated applications that receive direct information through the humanmachine connection, objective that has inputs and reduce the anomalies that occur in traditional methods, the high degree of inherent to the answers provided by user's freedom. Thus, there is interest in the development of systems to meet the emotional state of a user from the songs they listen, without it is subjected to the stress of performing a test or a physiological test that affects the outcome of your emotional state. Therefore, in this research project, the detection and classification of the emotional states of a user by using supervised learning techniques based on artificial intelligence was addressed, which allow meet the above difficulties. The results we obtained allow us to make a more accurate about the emotional state that presents the user, in time using the system, either short or long, with a simple and pleasant interaction trial. The main objective was the development of a methodology, based on an audio or multimedia system, detect the emotional state of the owner or failing of who uses it, as selected and played songs; why it has taken as its main theme the use of music as a means of detecting the emotional state, and thereby raise awareness of the importance of this research in providing future advances in detecting emotional states. The designed system was validated through a set of tests on simulation scenarios and case studies, which demonstrated excellent performance of the system, since in all tests detected correctly the emotional state from the playlist the songs heard. Another important contribution of this research is based on the creation of a (Dataset) for detecting emotional states from the music, which could be used to determine disorders or emotional disorders in users.
Descripción
Palabras clave
Estados de ánimo, Música, Detección emocional, Inteligencia artificial, Clasificación de datos, Aprendizaje supervisado, Moods, Music, Emotional detection, Artificial intelligence, Data classification, Supervised learning, Supervised learning
