We reduced the number of attributes by selecting courses that had high correlation with the programming course
CSC206. For English courses, we selected ENG121 and ENG122 for two main reasons. First, most students enrolled
in these courses, meaning that we had fewer NULL values. Second, they had a high correlation rate compared to
other English courses. We took the student’s grade from another English course or the average if a grade for one of
these courses was not available. In contrast, all math courses had the same correlation value, so we selected
MATH201, which was taught in Arabic, and MATH253, which was taught in English and had a high correlation
rate.