MOOC dropouts: A multi-system classifier
In recent years, technology enhanced learning platforms became widely accessible. In particular, the number of Massive Open Online Courses (MOOCs) has—and still is—constantly growing. This widespread adoption of MOOCs triggered the development of specialized solutions, that emphasize or enhance vari...
| Main Authors: | , , , , , |
|---|---|
| Format: | Conference Paper |
| Published: |
2017
|
| Online Access: | http://hdl.handle.net/20.500.11937/57314 |
| _version_ | 1848760049920376832 |
|---|---|
| author | Vitiello, M. Walk, S. Chang, V. Hernandez, R. Helic, D. Guetl, Christian |
| author_facet | Vitiello, M. Walk, S. Chang, V. Hernandez, R. Helic, D. Guetl, Christian |
| author_sort | Vitiello, M. |
| building | Curtin Institutional Repository |
| collection | Online Access |
| description | In recent years, technology enhanced learning platforms became widely accessible. In particular, the number of Massive Open Online Courses (MOOCs) has—and still is—constantly growing. This widespread adoption of MOOCs triggered the development of specialized solutions, that emphasize or enhance various aspects of traditional MOOCs. Despite this significant diversity in approaches to implementing MOOCs, many of the solutions share a plethora of common problems. For example, high dropout rate is an on-going problem that still needs to be tackled in the majority of MOOCs. In this paper, we set out to analyze dropout problem for a number of different systems with the goal of contributing to a better understanding of rules that govern how MOOCs in general and dropouts in particular evolve. To that end, we report on and analyze MOOCs from Universidad Galileo and Curtin University. First, we analyze the MOOCs of each system independently and then build a model and predict dropouts across the two systems. Finally, we identify and discuss features that best predict if users will drop out or continue and complete a MOOC using Boosted Decision Trees. The main contribution of this paper is a unified model, which allows for an early prediction of at-risk or dropout users across different systems. Furthermore, we also identify and discuss the most indicative features of our model. Our results indicate that users’ behaviors during the initial phase of MOOCs relate to their final results. |
| first_indexed | 2025-11-14T10:09:36Z |
| format | Conference Paper |
| id | curtin-20.500.11937-57314 |
| institution | Curtin University Malaysia |
| institution_category | Local University |
| last_indexed | 2025-11-14T10:09:36Z |
| publishDate | 2017 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | curtin-20.500.11937-573142018-01-10T05:54:44Z MOOC dropouts: A multi-system classifier Vitiello, M. Walk, S. Chang, V. Hernandez, R. Helic, D. Guetl, Christian In recent years, technology enhanced learning platforms became widely accessible. In particular, the number of Massive Open Online Courses (MOOCs) has—and still is—constantly growing. This widespread adoption of MOOCs triggered the development of specialized solutions, that emphasize or enhance various aspects of traditional MOOCs. Despite this significant diversity in approaches to implementing MOOCs, many of the solutions share a plethora of common problems. For example, high dropout rate is an on-going problem that still needs to be tackled in the majority of MOOCs. In this paper, we set out to analyze dropout problem for a number of different systems with the goal of contributing to a better understanding of rules that govern how MOOCs in general and dropouts in particular evolve. To that end, we report on and analyze MOOCs from Universidad Galileo and Curtin University. First, we analyze the MOOCs of each system independently and then build a model and predict dropouts across the two systems. Finally, we identify and discuss features that best predict if users will drop out or continue and complete a MOOC using Boosted Decision Trees. The main contribution of this paper is a unified model, which allows for an early prediction of at-risk or dropout users across different systems. Furthermore, we also identify and discuss the most indicative features of our model. Our results indicate that users’ behaviors during the initial phase of MOOCs relate to their final results. 2017 Conference Paper http://hdl.handle.net/20.500.11937/57314 10.1007/978-3-319-66610-5_22 restricted |
| spellingShingle | Vitiello, M. Walk, S. Chang, V. Hernandez, R. Helic, D. Guetl, Christian MOOC dropouts: A multi-system classifier |
| title | MOOC dropouts: A multi-system classifier |
| title_full | MOOC dropouts: A multi-system classifier |
| title_fullStr | MOOC dropouts: A multi-system classifier |
| title_full_unstemmed | MOOC dropouts: A multi-system classifier |
| title_short | MOOC dropouts: A multi-system classifier |
| title_sort | mooc dropouts: a multi-system classifier |
| url | http://hdl.handle.net/20.500.11937/57314 |