Analisis Pola Masa Studi Mahasiswa Fakultas Teknik Universitas Darma Persada Menggunakan Metode Clustering

Abstract

Penelitian ini bertujuan menganalisis pola lama studi mahasiswa fakultas teknik universitas Darma Persada daridata akademik. Metode yang digunakan adalah clustering algoritma K-Means. Variabel yang dianalisis adalah jurusan, daerah asal, umur, jenis kelamin, Indeks Prestasi Komulatif (IPK), Satuan Kredit Semester (SKS), tahun masuk, lama studi. Analisis dilakukan menggunakan perangkat lunak WEKA. Penelitian dilakukan melalui pengumpulan data dari arsip atau database biro Administrasi Akademik yaitu berupa data akademik mahasiswa fakultas teknik Universitas Darma Persada angkatan 2009 sampai 2014. Tahapan selanjutnya adalah preprocessing data yang dilakukan melalui analisis metode clustering menggunakan algoritma K-Means dengan terlebih dahulu menentukan jumlah cluster menggunakan metode Elbow dan interpretasi hasil. Berdasarkan hasil metode Elbow, jumlah cluster sebanyak 4 cluster. Berdasarkan hasil proses K-Means clustering, pembagian data pada masing-masing cluster adalah cluster 1 berjumlah 556 data (26%), cluster 2 berjumlah 414 data (19%), cluster 3 berjumlah 189 data (9%) dan cluster 4 berjumlah 1010 data (46%). Selanjutnya, yang memiliki lama studi lebih dari 4 tahun (lebih dari 8 semester) berada pada cluster 2, cluster 3, cluster 4 sedangkan mahasiswa yang memiliki masa studi 4 tahun (8 semester) berada pada cluster 1.AbstractThe duration of student study is one of the factors that influence the completing students' timeliness. Based on the policy of the National Accreditation Board of Higher Education (BAN-PT) in Regulation No. 4 of 2017 concerning the Policy for Preparing Accreditation Instruments, the duration of study is one of the benchmarks and evaluation elements in accreditation of study programs. From the Faculty of Engineering academic data, Darma Persada University, many students take more than four years of study. The duration of study is one of the problems of the study program manager in terms of academic performance. This study aims to analyze the old patterns of study by students of the Faculty of Engineering, Darma Persada University from academic data. K-Means algorithm clustering technique is used with the variables are majors, the area of origin, age, gender, Grade Point Average (GPA), Semester Credit Unit (SKS), year of entry and study duration. The Waikato Environment for Knowledge Analysis (WEKA) software is used as an analytic tool. The initial stage of research is through collecting data from archives or Academic sections, namely academic data from students of the Faculty of Engineering, Darma Persada University, 2009 to 2014. The next stage is preprocessing data through K-Means algorithm clustering analysis by first calculating many clusters using the Elbow method and result interpretation. From the Elbow method result, the number of clusters used is 4 (four) clusters. Based on the results of the K-Means clustering process, the data sharing in each cluster is cluster 1 (one) totaling 556 data (26%), cluster 2 (two) totaling 414 data (19%), cluster 3 (three) totaling 189 data (9%) and cluster 4 (four) totaling 1010 data (46%). Furthermore, those who have more than 4 years of study are in cluster 2, cluster 3, cluster 4 and students who have a 4-year study period are in cluster 1.

Keywords

• algoritma K-Means
• clustering
• data mining
• lama studi
• WEKA