OPTIMASI JUMLAH MAHASISWA ASIA MELALUI PREDIKSI MASA STUDI MENGGUNAKAN METODE INDUKSI DECISION TREE

Abstract

The accuracy of the study period can be determinant the students to pursued undergraduate degree. The available data indicated that only 30% of students can graduate <5 years, the rest graduate > 5 years and became non-active students. To overcome of low graduation rate is needed the system to determine the relationship between the master's students with study period that taken by the students. The use of data mining techniques in this system is expected to provide insights that were previously hidden in the data warehouse to be valuable information. By utilizing data mining techniques particularly algorithm ID3, then the researchers made an application to find a pattern that can predict the future of a student's study period based on the data from students and academic score. The student data, scores, and the study period integrated into the data training. The data training is processed into a decision tree based on the calculation of the gain and entropy. From that tree made a rule that can predict a student's study period. From 140 data training and 20 data testing with 6 kinds of attributes input and 2 kinds of target attributes, can be obtained by the accuracy of the ID3 prediction result for 85%, while the error rate prediction results for 15%.

Keywords : Optimasi, Jumlah Mahasiswa Asia, Prediks, Induksi Decision Tree