RANCANG BANGUN SISTEM PREDIKSI KELULUSAN MAHASISWA JURUSAN TEKNIK INFORMATIKA UNIVERSITAS PALANGKA RAYA

Abstract

The department is the leading part in the implementation of education from a college, so that it always conducts an evaluation to improve the quality and efficiency of higher education including
the improvement of graduate quality. The length of student study is one of the reference variables of
success level of the teaching process.
The graduation prediction system using the data mining classification method is Decisision
Tree C4.5. Data attributes used include; Gender, Religion, SKS, IPS, Graduated Semester, and TA
Type. The Graduated Semester attribute is used as a predictive target attribute. Where the attribute
value pass semester is made into 2 values that is 8-10 Semesters (<= 5 Years) and 11-14 Semesters
(> 5 Years). The prediction test was performed using k-fold cross-validation method and linear
regression measurement.
The highest accuracy score on the prediction system was obtained in the 6th experiment and
the 7th experiment was 61.54%. While for the lowest accuracy value obtained in the 5th experiment
of 30.77%. From the value of ????2 from experiment 1 to experiment 10 shows the highest value of 0.40
and the lowest 0.29. The value of ????2 obtained is so small that it can further explain the result of
prediction accuracy with decision tree C4.5 algorithm is very small value.