Exploring the Role of Small Differences in Predictive Accuracy using Simulated Data
| Authors | |
|---|---|
| Year of publication | 2015 |
| Type | Article in Proceedings |
| Conference | Proceedings of the Workshops at the 17th International Conference on Artificial Intelligence in Education |
| MU Faculty or unit | |
| Citation | |
| Field | Informatics |
| Keywords | student modeling; evaluation; simulated data; feedback |
| Description | Research in student modeling often leads to only small improvements in predictive accuracy of models. The importance of such improvements is often hard to assess and has been a frequent subject of discussions in student modeling community. In this work we use simulated students to study the role of small differences in predictive accuracy. We study the impact of such differences on behavior of adaptive educational systems and relation to interpretation of model parameters. We also point out a feedback loop between student models and data used for their evaluation and show how this feedback loop may mask important differences between models. |
| Related projects: |