Decision Tree Algorithm Use in Predicting Students’ Academic Performance in Advanced Programming Course

Abstract

Student’s academic performance or achievement has from time to time been a subject of discourse to academicians, scholars, researchers and educational institutions all over the globe. To this regard, schools are expected to play major and active roles in ensuring that students actually have good performance at end of their programmes. The academic performance is normally used to classify or predict how students would be ultimately capable to withstand and face future challenges after graduation. Students’ academic performance/achievement in any course of study plays a vital role in contributing and producing outstanding students who will be future viable leaders. The use of algorithms to classify and predict students’ academic performance/achievement is not new in machine learning using different techniques like neural network, logistic regression, decision tree and many more. This study classifies and predicts with the use of graphical technique called Decision Tree. The dataset was built from student’ attendance, practical assessment, assignment, ability to complete a free related course on internet, test score, and examination grade; the dataset was divided into training test and testing set. The training test was used to build and validate the decision tree algorithm (CHAID) while testing set was used to evaluate CHAID on the overall accuracy, sensitivity, and specificity. The results show that decision tree algorithm makes classification and prediction visible and clear with the use of graphics to display the results. Hence, the model built produces 96% accuracy.

Prediksi Kelulusan Siswa Sekolah Menengah Pertama Menggunakan Machine Learning

“…Berdasarkan beberapa penelitian yang telah dilakukan sebelumnya di dunia pendidikan, machine learning mampu melakukan tugas prediksi klasifikasi dengan baik. Algoritma decision tree diterapkan pada penelitian tentang memprediksi kinerja akademik [3], menilai atribut pembelajaran jarak jauh yang mempengaruhi performa mahasiswa [4], dan memprediksi kepuasan mahasiswa terhadap layanan akademik universitas [5]. Hasil yang diperoleh dalam penelitian ini menunjukkan hasil evaluasi yang baik dimana decision tree mampu menghasilkan nilai akurasi yang sangat baik.…”

“…Telah dilakukan beberapa penelitian menggunakan machine learning yang diterapkan dalam dunia pendidikan. Dilakukan klasifikasi dan prediksi performa akademik mahasiswa menggunakan data yang dibangun dari kehadiran mahasiswa, penilaian praktis, kumpulan nilai tugas, dan skor ujian mahasiswa [3]. Algoritma yang diterapkan adalah decision tree dan menghasilkan nilai evaluasi yang baik dengan nilai akurasi 96%.…”