Pemanfaatan Data PDDIKTI sebagai Pendukung Keputusan Manajemen Perguruan Tinggi

Abstract

<p>Pangkalan Data Pendidikan Tinggi (PDDIKTI) merupakan sebuah sistem penyimpan data yang dikelola Pusat Data dan Informasi (Pusdatin) Kementrian Ristek dan Pendidikan Tinggi. Data yang tersedia di PDDIKTI merupakan data yang akurat, karena proses pelaporan data akademik secara berkala dua kali setiap. Data yang telah berlimpah tersebut, tentu sangat disayangkan jika tidak digunakan untuk keperluan yang lebih bermanfaat, misal untuk mengetahui pola akademik kelulusan mahasiswa dan prestasi akademik mahasiswa. Untuk memperoleh informasi-informasi penting tersebut bisa dilakukan dengan cara penggalian informasi (<em>knowledge discovery</em>). Teknik dalam memberikan solusi masalah tersebut adalah teknik klasifikasi untuk membantu pengambilan keputusan, misalkan <em>Decission Tree</em> (C4.5, ID3, CHAID, <em>rule induction</em>) dan teknik peramalan (<em>forecasting</em>) menggunakan metode <em>simple moving average (SMA)</em>. Tujuan dari penambangan data PDDIKTI adalah untuk melakukan deteksi dini terhadap mahasiswa, sehingga dosen bisa memberikan masukan-masukan ketika mahasiswa tersebut telah diklasifikan sebagai mahasiswa yang lulus tidak tepat waktu serta memprediksi jumlah mahasiswa yang akan masuk pada perguran tinggi pada salah satu prodi X, sehingga manajemen baik tingkat program studi maupun universitas bisa melakukan langkah-langkah yang dianggap penting guna meningkatkan jumlah mahasiswa. Pengujian pada 2.601 <em>record</em> akademik mahasiswa dengan atribut ipk_sem1, ipk_sem2, ipk_sem3, ipk_sem4, pekerjaan_ortu, ket_lulus, rerata_ipk, penghasilan_ayah, untuk klasifikasi kelulusan mahasiswa menghasilkan nilai <em>accuracy</em> 86,54 % nilai <em>precission</em> 93,37% dan nilai <em>recall</em> 89,27% serta pengujian prediksi jumlah peminat program studi diperoleh nilai <em>accuracy</em> 78,25 % dan <em>MAPE</em> sebesar 21,75 %.</p><p class="Judul2"><em><br /></em></p><p class="Judul2"><em><strong>Abstract</strong> </em></p><p class="Judul2"><em><br /></em></p><p class="Judul2"><em>The Higher Education Database (PDDIKTI) is a data storage system managed by the Center for Data and Information (Pusdatin) of the Ministry of Research and Technology and Higher Education. The data available at PDDIKTI is accurate data, because the process of reporting academic data regularly twice each. The abundant data is certainly unfortunate if not used for more useful purposes, for example to find out the academic patterns of student graduation and student academic achievement. To obtain important information can be done by extracting information (knowledge discovery). Techniques in providing solutions to these problems are classification techniques to assist decision making, for example Decission Tree (C4.5, ID3, CHAID, rule induction) and forecasting techniques using simple moving average (SMA) methods. The purpose of PDDIKTI data mining is to conduct early detection of students, so that lecturers can provide input when the students have been classified as students who graduate not on time and predict the number of students who will enter the tertiary institutions in one of the X study programs, so that management both the level of study program and university can take steps that are considered important to increase the number of students. Tests on 2601 student academic records with the attributes ipk_sem1, ipk_sem2, ipk_sem3, ipk_sem4, occupation_ortu, graduated, average_ipk, income_ayah, for the graduation classification of students resulted in an accuracy value of 86.54% a value of 93.37% and a recall value of 89.27% and a test of 89.27% and a test of graduation prediction of the number of study program enthusiasts obtained an accuracy value of 78.25% and MAPE of 21.75%.</em></p>

Estimation of Gourami Supplies Using Gradient Boosting Decision Tree Method of XGBoost

“…Cases of data imbalance with the resulting information are usually referred to as the data-rich but information poor phenomenon where organizations have excess data from business processes that are captured and stored still lack of information [4]. Finding new information, patterns or certain rules in a large amount of data can be done using data mining techniques [5].…”