Collaborative recommendations in virtual learning environments.

González Bernal, Daniel

Resumen:

Recommender systems are software applications that aim to suggest meaningful and useful items to users chen interacting with large volumes of data such as online multimedia content, news, products, among others. These systems generate personalized recommendations of items based on the explicit or implicit preferences expressed by users, as well as information about their collaborations and relationships. there are numerous fields of application for these systems such as e-commerce, social networking, online streaming, among others, playing a key role in virtual learning environments. Recommender systems assist students in finding appropriate educational items, enhancing their learning experience and academia results. Two previous research projects (QHIR LACCIR and UTU) with limited scope, provided experience in different aspects of the development of recommendation algorithms which process information about socially-connected users and their collaborations. Based on those positive results obtained, this thesis proposes mechanisms to generate personalized recommendations to students through the development of a complete recommender system, using information about their collaborations in virtual learning environments. The aim is to describe and specify in detail wow different approaches, techniques, and recommendation algorithms are used to produce meaningful recommendations in real life educational scenarios. More specifically, it focuses on personalized recommendations for medical doctors and other health specialists enrolled in online Continuing Medical Education (CME) courses Iliroiigli a virtual learning environment end RedEMC. A general conclusion based on user feedback given by students, is that collaborative recommendations enhance each students’ learning process and user’s experience, stimulating further collaboration. It is worth to highlight that explicit feedback given by RedEMC’s users stated that more than 90% of the recommendations of resources/activities and comments were rated positively.


Los sistemas recomendadores son aplicaciones de software que tienen como objetivo sugerir elementos significativos y útiles para los usuarios cuando interactúan con grandes volúmenes de datos, como contenido multimedia en línea, noticias, productos, entre otros. Estos sistemas generan recomendaciones personalizadas de elementos basándose en las preferencias explícitas o implícitas expresadas por los usuarios, así como información sobre sus colaboraciones y relaciones. Existen numerosos campos de aplicación para estos sistemas, como el comercio electrónico, las redes sociales, la transmisión en línea, entre otros, desempeñando un papel clave en entornos de aprendizaje virtual. Los sistemas recomendadores ayudan a los estudiantes a encontrar elementos educativos apropiados, mejorando su experiencia de aprendizaje y resultados académicos. Dos proyectos de investigación anteriores (QHIR LACCIR y UTU) con alcance limitado, proporcionaron experiencia en diferentes aspectos del desarrollo de algoritmos de recomendación que procesan información sobre usuarios socialmente conectados y sus colaboraciones. Basados en estos resultados positivos obtenidos, esta tesis propone mecanismos para generar recomendaciones personalizadas para los estudiantes a través del desarrollo de un sistema recomendador completo, utilizando información sobre sus colaboraciones en entornos de aprendizaje virtual. El objetivo es describir y especificar en detalle cómo se pueden utilizar diferentes enfoques, técnicas y algoritmos de recomendación para producir recomendaciones significativas en escenarios educativos de la vida real. Más específicamente, se enfoca en recomendaciones personalizadas para médicos y otros especialistas de la salud inscriptos en cursos en línea de educación médica continua a través de un entorno de aprendizaje virtual denominado RedEMC. Una conclusión general basada en la retroalimentación dada por los usuarios es que las recomendaciones colaborativas en entornos virtuales de aprendizaje mejoran el proceso de aprendizaje de los estudiantes y la experiencia de usuaario de cada estudiante, estimulando una mayor colaboración. Cabe destacar que la retroalimentación explícita dada por los usuarios de RedEMC afirmaron que más del 90% de las recomendaciones de recursos/actividades y comentarios fueron calificadas positivamente.


Detalles Bibliográficos
2019
Sistemas recomendadores
Algoritmos de recomendación
Entornos de aprendizaje virtual
Inglés
Universidad de la República
COLIBRI
https://hdl.handle.net/20.500.12008/22855
Acceso abierto
Licencia Creative Commons Atribución - No Comercial - Sin Derivadas (CC - By-NC-ND 4.0)
Resumen:
Sumario:Recommender systems are software applications that aim to suggest meaningful and useful items to users chen interacting with large volumes of data such as online multimedia content, news, products, among others. These systems generate personalized recommendations of items based on the explicit or implicit preferences expressed by users, as well as information about their collaborations and relationships. there are numerous fields of application for these systems such as e-commerce, social networking, online streaming, among others, playing a key role in virtual learning environments. Recommender systems assist students in finding appropriate educational items, enhancing their learning experience and academia results. Two previous research projects (QHIR LACCIR and UTU) with limited scope, provided experience in different aspects of the development of recommendation algorithms which process information about socially-connected users and their collaborations. Based on those positive results obtained, this thesis proposes mechanisms to generate personalized recommendations to students through the development of a complete recommender system, using information about their collaborations in virtual learning environments. The aim is to describe and specify in detail wow different approaches, techniques, and recommendation algorithms are used to produce meaningful recommendations in real life educational scenarios. More specifically, it focuses on personalized recommendations for medical doctors and other health specialists enrolled in online Continuing Medical Education (CME) courses Iliroiigli a virtual learning environment end RedEMC. A general conclusion based on user feedback given by students, is that collaborative recommendations enhance each students’ learning process and user’s experience, stimulating further collaboration. It is worth to highlight that explicit feedback given by RedEMC’s users stated that more than 90% of the recommendations of resources/activities and comments were rated positively.