Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/146326
Title: CAERS: A Conversational Agent for Intervention in MOOCs' Learning Processes
Author: Rossi, Diego
Ströele, Victor
Braga, Regina
Caballé Llobet, Santi
Capuano, Nicola
Campos, Fernanda
Dantas, Mario
Lomasto, Luigi
Toti, Daniele
Others: Federal University of Juiz de Fora
Universitat Oberta de Catalunya
Università degli Studi della Basilicata
Università degli studi di Salerno
Università Cattolica del Sacro Cuore
Keywords: massive open online courses
recommender system
conversational agent
Issue Date: 13-Nov-2021
Publisher: Lecture Notes in Networks and Systems
Citation: Rossi, D., Ströele, V., Braga, R., Caballé, S., Capuano, N., Campos, F., Dantas, M., Lomasto, L., & Toti, D. (2022). CAERS: A Conversational Agent for Intervention in MOOCs' Learning Processes. In D. Guralnick, M. E. Auer, & A. Poce (Eds.), Innovations in Learning and Technology for the Workplace and Higher Education (Vol. 349, pp. 371-382). Springer International Publishing. doi: 10.1007/978-3-030-90677-1_36
Published in: 349;
Also see: https://link.springer.com/chapter/10.1007/978-3-030-90677-1_36
Abstract: Massive Open Online Courses (MOOCs) make up a teaching modality that aims to reach a large number of students using Virtual Learning Environments. In these courses, the intervention of tutors and teachers is essential to support students in the teaching-learning process, answer questions about their content, and provide engagement for students. However, as these courses have a vast and diverse audience, tutors and teachers find it difficult to monitor them closely and efficiently with prompt interventions. This work proposes an architecture to favor the construction of knowledge for students, tutors, and teachers through autonomous interference and recommendations of educational resources. The architecture is based on a conversational agent and an educational recommendation system. For the training of predictive models and extraction of semantic information, ontology and logical rules were used, together with inference algorithms and machine learning techniques, which act on a dataset with messages exchanged between course forum participants in the humanities, medicine, and education fields. The messages are classified according to the type (question, answer, and opinion) and parameters about feeling, confusion, and urgency. The architecture can infer the moment in which a student needs help and, through a Conversational Recommendation System, provides the student with the opportunity to revise his or her knowledge on the subject. To help in this task, the architecture can provide educational resources via an autonomous agent, contributing to reducing the degree of confusion and urgency identified in the posts. Initial results indicate that integrating technologies and resources, complementing each other, can support the students and help them succeed in their educational training.
Language: English
URI: http://hdl.handle.net/10609/146326
ISBN: 978-3-030-90676-4
Appears in Collections:Conference lectures

Files in This Item:
File Description SizeFormat 
rossi_lnns_caers.pdf514,62 kBAdobe PDFThumbnail
View/Open