1. DSpace Universitas Potensi Utama
  2. Jurnal Nasional dan Terakreditasi P-B
  3. Untitled
Please use this identifier to cite or link to this item: http://repository.potensi-utama.ac.id/jspui/jspui/handle/123456789/5463
Title: Students Grade Grouping to Optimize On-Time Graduation Predictions by Combining K-Means and C4.5 Algorithms (Case Study : University Potensi Utama
Authors: Bob Subhan, Riza
Keywords: K-Means, C4.5, Graduated on time, Student Grades
Issue Date: 20-Feb-2021
Publisher: Telematika
Series/Report no.: Vol. 14, No. 1;
Abstract: Graduating on time is the dream of every student who studies in universities. Some factors that can lead to failure in graduating on time, such as grades, though students are sometimes careless and underestimating this factor, despite knowing that problematic Grade will hinder the student from graduating on time. This research helps the study program to predict which students will graduate on time. There are 2 stages in the research, first is the process of clustering students' data using the KMeans algorithm, while the second stage predicts students' graduation using the C4.5 algorithm. Variable used are Grade, Failing Grade, Specialization, Internship, Thesis, Undergraduate Thesis 1, Undergraduate Thesis 2, and Passing Grade. Using RapidMiner and processing these data using this software can predict students that graduate on time.
Description: The problem that occurs today is students who do not realize nor care about the problematic grades (failed Grade). Problematic grades can affect students in taking specialization, street vendors, thesis 1 and thesis 2 not on the right time, where it impacts graduation. In the 2015 information system study program of students' data showed that there were 0.49% of students who had failed or problematic grades that did not pass on time, and so were students who could not complete internship and thesis on time, thus resulting in graduating not on time with a percentage value of 100%. So the criteria that will be included in this case study are Failing Grade, Grade, Specialization, Internship, Undergraduate Thesis 1 and 2. This case study also uses RapidMiner software. The clustering method uses K-Means, which functions to cluster students' data to make it easier for the next process, and C4.5 is used to process predictive decisions of students who can graduate on time. This research only discusses the prediction of students graduating on time, where the benefits are that students are more concerned with grades and courses to be taken next to graduate in time, and the study program can find out which students have problematic grades.
URI: http://repository.potensi-utama.ac.id/jspui/jspui/handle/123456789/5463
ISSN: 2442-4528
Appears in Collections:Untitled

Files in This Item:
File Description SizeFormat 
10. SINTA 2_Students Grade Grouping to Optimize On-Time Graduation Predictions.pdf619.04 kBAdobe PDFView/Open
Show full item record


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.