Publicación: Análisis de entornos colaborativos con tecnologías de análisis de redes y Learning Analytics
Fecha
2023-09
Autores
Editor/a
Director/a
Tutor/a
Coordinador/a
Prologuista
Revisor/a
Ilustrador/a
Derechos de acceso
info:eu-repo/semantics/openAccess
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Nacional de Educación a Distancia (España). Escuela Técnica Superior de Ingeniería Informática
Resumen
El presente Trabajo Fin de Máster consiste en la realización y prueba de una aplicación que permita el análisis de sentimientos y emociones, así como el análisis de la red social subyacente de los foros de los cursos virtuales de los diferentes estudios que se pueden cursar en la UNED y por extensión se puede utilizar para analizar otros foros o cualesquiera colecciones de mensajes o textos. Sobre la base teórica y práctica de trabajos previos se ha construido una aplicación que pretende ser más amigable y fácil de usar y con un potencial uso práctico y de investigación. Este trabajo se apoya en dos trabajos previos dentro del ámbito de la UNED, como son otro trabajo de fin de máster de la alumna Andrea Rey Presas y la aplicación web appForum , presentada en las XI Jornadas de Investigación en Innovación Docente de la UNED, si bien es importante reseñar que esta es una aplicación totalmente distinta a cualesquiera otras, tanto en lo que se refiere a la filosofía como a la funcionalidad, presentación, etc. Lo que aquí nos encontramos es una aplicación de escritorio con una interfaz gráfica basada en Tkinter (Python). Se organiza en diferentes módulos funcionales, destacando el módulo de modelos e informes, el cual permite obtener un resumen de los análisis realizados y generar un informe automático en formato PDF. De los mencionados trabajos se han aprovechado algunas estructuras algorítmicas de los mismos, se ha extendido la funcionalidad, se ha creado una interfaz lo más amigable posible y se ha dotado de flexibilidad para poder trabajar con diferentes tipos de archivos, así como configuraciones de los mismos. Se propone un método alternativo para el descubrimiento de sentimientos de los mensajes. Dicho método consiste en la combinación lineal de varios métodos, cuya fórmula es configurable. Respecto del análisis de emociones se realizan dos análisis diferentes que pueden compararse. Un primer análisis basado en regiones de Russell a partir de un corpus de términos etiquetados con valores de valencia y excitación. El segundo método consiste también en la utilización de un corpus de términos, pero en este caso se les han asignado puntuaciones para cada una de varias emociones consideradas primarias. Este segundo método se ha probado con el corpus de 2.226 palabras puntuadas para cinco emociones primarias creado por (Ferré et al.) en el trabajo Moved by words: Affective ratings for a set of 2,266 Spanish words in five discrete emotion categories. La aplicación se ha probado con datos reales de foros de la UNED anonimizados y se ha utilizado como herramienta para analizar sentimientos en otro trabajo fin de máster: Estudio sobre la búsqueda de mentor utilizando aprendizaje automático realizado por D. Juan García Ruiz y dirigido también por el Dr. Antonio Rodríguez Anaya.
This final master's project consists of the development and testing of an application to carry out both the analysis of feelings and emotions of the messages sent in the forums of the UNED university, as well as the analysis of the social network created in said forums. By extension it can be used to analyze other forums or any collections of messages or texts. On the theoretical and practical basis of previous work, an application has been built that aims to be friendlier and easier to use and with potential practical and research use. This work is based on two previous works, such as another master's thesis by the student Andrea Rey Presas and a web application (appForum ), presented at the XI Conference on Research in Teaching Innovation of the UNED, although it is important to note that this is an application totally different from any other, both in terms of approach and functionality, presentation, etc. What we find here is a desktop application with a graphical interface based on Tkinter (Python). It is organized into different functional modules, highlighting the models and reports module, which allows obtaining a summary of the analyzes carried out and generating an automatic report in PDF format. Some of the code structures of the works mentioned above have been used and the functionality has been extended, an interface that is as friendly as possible has been created and flexibility has been provided to be able to work with different types of files, as well as their configurations. An alternative method is proposed for the discovery of message sentiments based on the linear combination of several methods. The formula that defines this method can be configured. Regarding the analysis of emotions, two different analyzes are carried out that can be compared. A first analysis based on Russell regions from a corpus of terms labeled with valence and excitation values. The second method also consists of using a corpus of terms, but in this case scores have been assigned for each of several emotions considered primary. This second method has been tested with the corpus of 2,226 words scored for five primary emotions created by (Ferré et al.) in the work entitled Moved by words: Affective ratings for a set of 2,266 Spanish words in five discrete emotion categories. The application has been tested with real data from anonymized UNED forums and it has been used as a tool to sentiment analysis in another master's thesis: Study on the mentor searching using machine learning carried out by Juan García Ruiz and also directed by Dr. Antonio Rodríguez Anaya.
This final master's project consists of the development and testing of an application to carry out both the analysis of feelings and emotions of the messages sent in the forums of the UNED university, as well as the analysis of the social network created in said forums. By extension it can be used to analyze other forums or any collections of messages or texts. On the theoretical and practical basis of previous work, an application has been built that aims to be friendlier and easier to use and with potential practical and research use. This work is based on two previous works, such as another master's thesis by the student Andrea Rey Presas and a web application (appForum ), presented at the XI Conference on Research in Teaching Innovation of the UNED, although it is important to note that this is an application totally different from any other, both in terms of approach and functionality, presentation, etc. What we find here is a desktop application with a graphical interface based on Tkinter (Python). It is organized into different functional modules, highlighting the models and reports module, which allows obtaining a summary of the analyzes carried out and generating an automatic report in PDF format. Some of the code structures of the works mentioned above have been used and the functionality has been extended, an interface that is as friendly as possible has been created and flexibility has been provided to be able to work with different types of files, as well as their configurations. An alternative method is proposed for the discovery of message sentiments based on the linear combination of several methods. The formula that defines this method can be configured. Regarding the analysis of emotions, two different analyzes are carried out that can be compared. A first analysis based on Russell regions from a corpus of terms labeled with valence and excitation values. The second method also consists of using a corpus of terms, but in this case scores have been assigned for each of several emotions considered primary. This second method has been tested with the corpus of 2,226 words scored for five primary emotions created by (Ferré et al.) in the work entitled Moved by words: Affective ratings for a set of 2,266 Spanish words in five discrete emotion categories. The application has been tested with real data from anonymized UNED forums and it has been used as a tool to sentiment analysis in another master's thesis: Study on the mentor searching using machine learning carried out by Juan García Ruiz and also directed by Dr. Antonio Rodríguez Anaya.
Descripción
Categorías UNESCO
Palabras clave
Social Network Analysis, Sentiment Analysis, Discrete emotion categories, Affective norms, Emotional effects on word processin, Learning Analytics
Citación
Centro
Facultades y escuelas::E.T.S. de Ingeniería Informática
Departamento
No procede