Modeling classifiers for virtual internships without participant data
Virtual internships are online simulations of professional practice where students play the role of interns at a fictional company. During virtual internships, participants complete activities and then submit write-ups in the form of short answers, digital notebook entries. Prior work used classifiers trained on participant data to automatically assess notebook entries from these learning environments. However, when teachers create new internships using available authoring tools, no such data exists. We evaluate a method for generating classifiers using specifications provided by teachers during their authoring process instead of participant data. Our models rely on Latent Semantic Analysis based and Neural Network based semantic similarity approaches in which notebook entries are compared to ideal, expert generated responses. We also investigated a Regular Expression based model. The experiments on the proposed models on unseen data showed high precision and recall values for some classifiers using a similarity based approach. Regular Expression based classifiers performed better where the other two approaches did not, suggesting that these approaches may complement one another in future work.
Proceedings of the 10th International Conference on Educational Data Mining, EDM 2017
Gautam, D., Swiecki, Z., Shaffer, D., Graesser, A., & Rus, V. (2017). Modeling classifiers for virtual internships without participant data. Proceedings of the 10th International Conference on Educational Data Mining, EDM 2017, 278-283. Retrieved from https://digitalcommons.memphis.edu/facpubs/2972