Open Universiteit

Please use this identifier to cite or link to this item:
Title: Automated Prediction of Student Participation in Collaborative Dialogs Using Time Series Analyses
Authors: Pasov, Iulia
Dascalu, Mihai
Nistor, Nicolae
Trausan-Matu, Stefan
Keywords: CSCL
Time series analysis
Automated evaluation of participation
Issue Date: Jun-2018
Publisher: Springer
Citation: Pasov, I., Dascalu, M., Nistor, N., & Trausan-Matu, S. (2018). Automated Prediction of Student Participation in Collaborative Dialogs using Time Series Analyses. In H. Knoche, E. Popescu & A. Cartelli (Eds.), 3rd Int. Conf. on Smart Learning Ecosystems and Regional Development (SLERD 2018) (pp. 177–185). Aalborg, Denmark: Springer.
Abstract: The massive student participation in Computer Supported Collaborative Learning (CSCL) sessions from online classrooms requires intense tutor engagement to track and evaluate individual student participation. In this study, we investigate how the time evolution of messages predicts students’ participation using two models – a linear regression and a Random Forest model. A corpus of 10 chats involving 47 students was scored by 4 human experts and used to evaluate our models. Our analysis shows that students’ pauses length between consecutive messages within a discussion is the strongest participation predictor accounting for R2 ¼ :796 variance in the human estimations while using a Random Forest model. Our results provide an extended basis for the automated assessment of student participation in collaborative online discussions.
Appears in Collections:1. RAGE Publications

Files in This Item:
File Description SizeFormat 
SLERD Iulia.pdf257.87 kBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons