Drop Columns: Curricular units 1st sem and 2nd sem (credited), (Evaluations), ...
Without Evaluations are extra curricular activites like projects, workshops non graded
Admission Grade is between 0 and 200
The Portuguese grading system is based on a 0 to 20 scale (corresponding ECTS grade within brackets): Less than 10 : FAIL (F) 10: Minimum pass grade (E) 11 - 12: SUFFICIENT (D) 13 - 14: GOOD (C) 15 – 17: VERY GOOD (B) 18 - 20: EXCELLENT (A)
Working with Enrolled Labels: We gonna judge them on their grades (1st and 2nd Sem) and sort them accordingly then we gonna label appprox 40% them as droppers.
Dropout: 0 Graduated: 1
We remove enrolled students with 0 credits in 2nd semester and 1st Sem both then we remove them
Map Nationalities to 1st,2nd,3rd World countries
People who dropped out after 1st sem
Convert grades from random scale to percentages
we train the model only on the dropouts (1421) because graduates weaken the model.
We dont use 2nd Sem grade anymore except enrolled column, because if the person didnt enroll in 2nd semester and dropped out we consider they droppedd out in first semester. Rest dropped out in higher semester.
So Model 1: We dont use university data and judge the student whether he graduated or droppedd on the basis of past qualifications and family background
Model 2: We use only dropuots to figure out whether they dropped in 1st sem or in higher sem.
##Relevant Wesites:
https://archive.ics.uci.edu/dataset/697/predict+students+dropout+and+academic+success https://www.mdpi.com/2306-5729/7/11/146#fig_body_display_data-07-00146-f002 https://www.mdpi.com/2306-5729/7/11/146 https://github.com/carmelh/SQL_projects/tree/main/student_data_analysis/Datasets https://www.kaggle.com/datasets/thedevastator/higher-education-predictors-of-student-retention/discussion/392836