Co-training algorithm on Educational Data Mining
An implementation of the Co-training scheme, a well-known multi-view Semi-Supervised Learning approach, applied on Educational Data Mining datasets related with the task of Early Prognosis of Academic Performance is provided here.
Apart from the proposed algorithm Cotrain(Extra, GBC), the results of several other variant of Co-training scheme are provided, as well as the results of Self-training approaches. Moreover, the results of the CoForest algorithm (implemented by Mr. Ming Li email@example.com - link) have been computed for the same datasets' splits and same seeds.
A full description is provided on the related publication. More comments are going to be posted, after the acceptance of the submission.
Please cite this paper if you use our algorithm/datasets for your work.
Early access: https://ieeexplore.ieee.org/document/8692618
Python dependencies for main algorithm
pip install -r requirements.txt
Python dependencies for visualizations algorithm
pip install -r requirements_draw.txt
More information about the authors are provided in ml.math.upatras.gr