Pass/Fail Prediction in Programming Courses

We present a privacy-friendly early-detection framework to identify students at risk of failing in introductory programming courses university. The was validated for two different with annual editions taken by higher education ( N = 2 080) and found be highly accurate robust against variation course structures, teaching learning styles, exercises classification algorithms. By using interpretable machine techniques, the also provides insight into what aspects practising skills promote or inhibit have no minor effect on process. Findings showed that capable predicting students’ future success already early semester.