Concept

Discussion

As it was found, homework assignments, demo exams, prior programming knowledge are correlated with final exam grade. Still, individual characteristics aren't considered in these variables that can explain low accuracy of the model 52%. Also, here as the authors have explained, they have multicollinearity issue. Lack of the features can also be the reason for having low performance. In case of predicting students who are at risk, there was even lower performance(46 4).

0

1

Updated 2021-03-02

Tags

Data Science