Publication: Análisis de sentimiento en audio mediante inteligencia artificial orientado al idioma español
Loading...
Identifiers
Publication date
2022
Defense date
2022-07
Authors
Advisors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
En este proyecto se presenta un trabajo como proyecto de investigación de la cátedra entre
la Universidad Carlos III de Madrid y el Grupo Masmovil.
El objetivo es crear una Inteligencia Artificial capaz de clasificar la emoción de los
clientes en las llamadas dentro del Call Center. Pero la naturaleza agnóstica del problema permite enfocarlo desde una perspectiva más amplia, a partir de la base de clasificar
emociones en audio para el idioma Español, es posible adaptar este trabajo a diferentes
propósitos.
A partir de los principios del reconocimiento de emociones en el habla, se estudian
distintas aproximaciones para transmitir el conocimiento de lo que representa una emoción a un modelo matemático. Además, se combinan los resultados con el sentimiento
que transmite el aspecto textual haciendo uso de algoritmos de aprendizaje automático.
Además, este trabajo viene integrado en una interfaz de una aplicación de mensajería
para probarlo en tiempo real.
This present document portraits a project carried out as part of the chair between the Carlos III University of Madrid and the Masmovil Group. The proposed objective is to create an Artificial Intelligence capable of classifying the emotion of customers in calls inside the Call Center. The agnostic nature of the problem allows to approach it from a broader perspective, from the basis of classifying emotions in audio for the Spanish language, it is possible to adapt this work for different purposes. Based on the principles of emotion recognition in speech, different approaches are studied to convey the knowledge of what an emotion represents to a mathematical model. In addition, the results are combined with the sentiment conveyed by the textual aspect using machine learning algorithms. Besides, this work is integrated in a messaging application interface to test it in real time.
This present document portraits a project carried out as part of the chair between the Carlos III University of Madrid and the Masmovil Group. The proposed objective is to create an Artificial Intelligence capable of classifying the emotion of customers in calls inside the Call Center. The agnostic nature of the problem allows to approach it from a broader perspective, from the basis of classifying emotions in audio for the Spanish language, it is possible to adapt this work for different purposes. Based on the principles of emotion recognition in speech, different approaches are studied to convey the knowledge of what an emotion represents to a mathematical model. In addition, the results are combined with the sentiment conveyed by the textual aspect using machine learning algorithms. Besides, this work is integrated in a messaging application interface to test it in real time.
Description
Keywords
Inteligencia artificial, Aprendizaje automático, Procesamiento del lenguaje natural