Regresi logistik multinomial untuk prediksi kategori kelulusan mahasiswa

Suhartono, Suhartono, Syahranita, Rafika and Zaman, Syahiduz (2023) Regresi logistik multinomial untuk prediksi kategori kelulusan mahasiswa. JISKA (Jurnal Informatika Sunan Kalijaga), 8 (2). ISSN 2527 –5836

Text 20588.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (281kB)

Abstract

Students must meet certain goals to earn a degree but can extend their time at university or drop out (DO). The problem of dropping out of students has become an important issue for tertiary institutions to ensure the success or graduation of students and reduce dropouts. DOcan affect the accreditation of the tertiary institution. The quality of higher education institutions in Indonesia is measured based on accreditation from the National Accreditation Board for Higher Education or BAN-PT. One of the main standards measuredis the Quality of Students and Graduates. The quality of educational accreditation is measured by the percentage of student graduation and the university's strategy to retain students. To predict student graduation based on graduation time categories, researchers collected academic data from students in 2012-2018 at the Informatics Engineering Study Program, State Islamic University of Maulana Malik Ibrahim Malang. The variables used as predictors are gender, type of entry pathway, and grade point average from semesters one to six. The resulting model was evaluated to obtain an accuracy value of 85.5%, a precision of 78.5%, a recall of 93.9%, and a micro f1-score of 89.8%. An accuracy value of 85.5% indicates that the system can classify properly using the logistic regression model.

Item Type:	Journal Article
Keywords:	Categories; Graduation; Prediction; Logistic Regression; Machine Learning
Subjects:	08 INFORMATION AND COMPUTING SCIENCES > 0801 Artificial Intelligence and Image Processing > 080102 Artificial Life
Divisions:	Faculty of Technology > Department of Informatics Engineering
Depositing User:	Dr Suhartono M.Kom
Date Deposited:	14 Oct 2024 15:23

Downloads

Origin of downloads

Actions (login required)

View Item