- https://link.springer.com/chapter/10.1007%2F978-3-030-49663-0_10
- https://its2020.iis-international.org/wp-content/uploads/1.-18-SHAPed-Automated-Essay-Scoring_-Explaining-Writing-Features%E2%80%99-Contributions-to-English-Writing-Organization.pdf
### Background ###
In a previous paper, we developed an explanation model for a multioutput classifier AES model and analyzed the decision process behind the predicted scores of the Ideas rubric. Moreover, it was found that the feature selection process was not optimal, even potentially affecting a bit the generalizability of the AES. This paper addresses the issues related to the feature selection process, trains an ensemble regression AES model, and analyzes the reasoning of the AES system for the organization rubric.
Based on https://github.com/slundberg/shap
Published in the setting of the 2020 International Conference on Intelligent Tutoring Systems.
