Exploring Automatic Assessment-Based Features for Clustering of Students’ Academic Performance

Many students have difficulty learning computer programming languages. Programming is a skill that requires a lot of practice, not just theory. Students are required to have the ability to all processes; analyze problems, design algorithms, translate algorithms into program code, and write program code with the correct syntax. The number of time students spend coding is the factor that has the highest impact on their programming skills. A basic programming lab work is implemented to increase students' programming skills with a lot of practice writing a code directly. Providing feedback on programming assignments is an integral part of a class on basic programming and requires substantial effort with personal teaching. It is needed an automatic assessment tool that can help the task of lecturers in evaluating assignments. Performance evaluation is one of the basics to monitor the progress of student performance. Grouping students according to their level of performance makes it easy for lecturers to monitor student performance levels and can provide learning according to the abilities of students in these groups. The clustering is used to group students based on the point of each lab work. The classification method used is K-Means Clustering. The cluster describes groups of students according to their performance. Based on the results of a case study with 31 students, it clusters students into 3 groups: 39% of people are in moderate ability, 45% of people have high abilities, and 16% of people are students whose programming abilities are still lacking system.