| Full text | |
| Author(s): Show less - |
Penteado, Bruno Elias
;
Isotani, Seiji
;
Paiva, Paula M.
;
Morettin-Zupelari, Marina
;
Ferrari, Deborah Viviane
;
Rose, CP
;
Martinez-Maldonado, R
;
Hoppe, HU
;
Luckin, R
;
Mavrikis, M
;
Porayska-Pomsta, K
;
McLaren, B
;
DuBoulay, B
Total Authors: 13
|
| Document type: | Journal article |
| Source: | ARTIFICIAL INTELLIGENCE IN EDUCATION, PT II; v. 10948, p. 5-pg., 2018-01-01. |
| Abstract | |
We propose a machine learning approach to automate the estimation of the interpersonal help-seeking level of students in an online course, based on their behavior in an LMS platform. We selected behavioral and performance features from the LMS logs, using forum and wiki variables in the context of a professional development course in audiology rehabilitation (N = 93). Then, we applied different state-of-the-art regression algorithms to predict their responses, using student-level cross-validation in the training set and evaluated the resulting models in a separate test set. As result, we had approximately an error of one point with our model, on average. We discuss some deviant cases and how this information can be used to inform tutors in online courses. (AU) | |
| FAPESP's process: | 15/24507-2 - Ecosystem for production and consumption of connected open data and its application in educational settings |
| Grantee: | Seiji Isotani |
| Support Opportunities: | Regular Research Grants |