Penerapan Adaboost Berbasis Pohon Keputusan Guna Menentukan Pola Masuknya Calon Mahasiswa Baru

Abstract

In general, the college admission process is done through registration, file selection, examinations, an announcement of the results of students who pass, and ends with re-registration. In this case, a problem was found where there is a significant decrease in the number of student who register with those who re-register .Things like this can reduce the balance between new students and students who meet the requirements, to make a decrease in the quality of higher education and affect accreditation. Based on these problems, a classification method was developed to look for patterns of students who would enter institutions and what factors influence students to re-register.

To improve the accuracy of the decision tree algorithm the author use adaptive boosting (adaboost) in finding factors that make prospective students continue to the re-registration process.

From the results of the study, the AdaBoost-based decision tree algorithm shows that the level of accuracy has an increase of 20%. The presentation of results is as follows, 61.4% (decision tree); 91.35% (decision tree + AdaBoost)