Please use this identifier to cite or link to this item:
https://repository.isls.org/handle/1/10605| Title: | An Application of Data Mining Methods on In-Game Behaviors: Predicting Student Math Performance |
| Authors: | Pradhan, Siddhartha Lee, Ji-Eun Egorova, Alena Jerusal, Janette Ottmar, Erin |
| Keywords: | Learning Sciences |
| Issue Date: | 2024 |
| Publisher: | International Society of the Learning Sciences |
| Citation: | Pradhan, S., Lee, J., Egorova, A., Jerusal, J., & Ottmar, E. (2024). An Application of Data Mining Methods on In-Game Behaviors: Predicting Student Math Performance. In Lindgren, R., Asino, T. I., Kyza, E. A., Looi, C. K., Keifert, D. T., & Suárez, E. (Eds.), Proceedings of the 18th International Conference of the Learning Sciences - ICLS 2024 (pp. 1007-1010). International Society of the Learning Sciences. |
| Abstract: | This study tested the applicability of the Random Forest (RF) prediction model built in Lee et al. (2023) to a different sample with a larger number of students (N = 681). Specifically, we predicted middle school students’ math knowledge scores using both in-game behavioral features in a digital algebraic learning game and math anxiety scores. Further, we conducted a Principal Component Analysis (PCA) to test whether dimensionality reduction improved the performance of the prediction model. The results showed that the RF regressor performed well in predicting posttest math knowledge scores even when trained on data without pretest scores. Out of the 20 features included in the prediction model, the average number of steps a student takes in their first attempt was the most predictive of their posttest scores. Finally, using PCA as preprocessing to minimize multicollinearity and noise improved the model performance. |
| Description: | Short Paper |
| URI: | https://doi.org/10.22318/icls2024.201482 https://repository.isls.org//handle/1/10605 |
| Appears in Collections: | ISLS Annual Meeting 2024 |
Files in This Item:
| File | Size | Format | |
|---|---|---|---|
| ICLS2024_1007-1010.pdf | 368.66 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.