Article Per Year (5 Year)
p-Index From 2021 - 2026
1.827
P-Index
This Author published in this journals
All Journal JSiI (Jurnal Sistem Informasi) Jurdimas (Jurnal Pengabdian Kepada Masyarakat) Royal JOURNAL OF SCIENCE AND SOCIAL RESEARCH BIOS : Jurnal Teknologi Informasi dan Rekayasa Komputer International Journal of Management Science and Information Technology (IJMSIT) Journal Of Indonesian Social Society
Cecep Maulana
Universitas Royal
Religion Humanities Computer Science & IT Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Languange, Linguistic, Communication & Media Law, Crime, Criminology & Criminal Justice Social Sciences Other

Published : 12 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 12 Documents
Search

 1   2 
KLASIFIKASI KESESUAIAN BIDANG PELATIHAN CALON TENAGA KERJA MENGGUNAKAN INFORMATION GAIN DAN RANDOM FOREST Rika Nofitri; Suparmadi Suparmadi; Cecep Maulana
JOURNAL OF SCIENCE AND SOCIAL RESEARCH Vol. 9 No. 4 (2026): August 2026 (1)
Publisher : Smart Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54314/jssr.v9i4.7114

Abstract

Abstract: Selecting appropriate vocational training fields for prospective workers at Job Training Centers (BLK) is critical for optimizing skill development and ensuring job market alignment. However, manual selection processes and high-dimensional datasets containing irrelevant candidate attributes often lead to classification inaccuracies and computational noise. This study proposes an optimized classification model combining the Information Gain (IG) feature selection method with the Random Forest algorithm to enhance prediction accuracy for training field suitability. Using a dataset of 233 candidate records, feature weighting was executed via Information Gain to eliminate redundant demographic attributes. The feature selection identified four dominant predictors: Academic Potential Test (TPA) score, interview score, vocational skill score, and BLK training choice. Experimental evaluation was conducted using a 70:30 train-test split in RapidMiner Studio. The experimental results demonstrate that applying Information Gain reduced the feature space from nine to four attributes, increasing classification accuracy from 87.14% (baseline Random Forest) to 90.00% (+2.86%). Furthermore, the model achieved a notable precision enhancement, rising from 86.96% to 93.44% (+6.48%), and an improved F1-Score of 94.21%. These findings confirm that removing non-contributing demographic variables mitigates dataset noise and enables Random Forest to form more effective decision boundaries. The optimized model offers a robust, objective framework to support decision-making systems in candidate placement at vocational training institutions. Keywords: Information Gain, Random Forest, Feature Selection, Training Field Classification, RapidMiner.   Abstrak: Penentuan bidang pelatihan vokasi yang tepat bagi calon peserta di Balai Latihan Kerja (BLK) sangat krusial untuk mengoptimalkan pengembangan keterampilan dan kesesuaian dengan kebutuhan pasar kerja. Namun, proses seleksi manual dan tingginya dimensi dataset yang memuat atribut kandidat kurang relevan sering kali menyebabkan ketidakakuratan klasifikasi serta noise komputasi. Penelitian ini mengusulkan model klasifikasi teroptimasi yang menggabungkan metode seleksi fitur Information Gain (IG) dengan algoritma Random Forest untuk meningkatkan akurasi prediksi kesesuaian bidang pelatihan. Menggunakan dataset sebanyak 233 data calon peserta, pembobotan fitur dieksekusi melalui Information Gain untuk mengeliminasi atribut demografis yang redundan. Hasil seleksi fitur berhasil mengidentifikasi empat prediktor dominan: nilai Tes Potensi Akademik (TPA), nilai wawancara, nilai kejuruan, dan pilihan BLK. Evaluasi eksperimental dilakukan dengan pembagian data train-test sebesar 70:30 pada perangkat lunak RapidMiner Studio. Hasil pengujian membuktikan bahwa penerapan Information Gain yang mereduksi ruang fitur dari sembilan menjadi empat atribut mampu meningkatkan akurasi klasifikasi dari 87,14% (model awal Random Forest) menjadi 90,00% (+2,86%). Selain itu, model ini mencapai peningkatan presisi yang signifikan, naik dari 86,96% menjadi 93,44% (+6,48%), serta kenaikan F1-Score menjadi 94,21%. Temuan ini mengonfirmasi bahwa penghapusan variabel demografis yang tidak berkontribusi berhasil mengurangi noise data dan memungkinkan Random Forest membentuk pohon keputusan yang lebih efektif. Model teroptimasi ini memberikan kerangka kerja yang objektif dan andal untuk mendukung sistem pengambilan keputusan dalam penempatan calon peserta di lembaga pelatihan vokasi. Kata kunci: Information Gain, Random Forest, Seleksi Fitur, Klasifikasi Bidang Pelatihan, RapidMiner.
Penerapan Fitur PBL E-CRM untuk Meningkatkan User Engagement di Aditya Kursus Putri Oktaviya Siregar; Nasrun Marpaung; Cecep Maulana
BIOS : Jurnal Teknologi Informasi dan Rekayasa Komputer Vol 7 No 2 (2026): September (In Progress)
Publisher : Puslitbang Sinergis Asa Professional

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37148/bios.v7i2.242

Abstract

Advancements in information technology have driven the adoption of digital systems across various sectors including non-formal education to support service management and participant relations. This study aims to design and develop a web-based Electronic Customer Relationship Management (E-CRM) system for Aditya Kursus, capable of integrating participant management and service delivery in a more structured manner. A qualitative approach was employed, utilizing data collection methods such as observation, interviews, and literature reviews, while analysis involved interpreting information from diverse sources to determine system requirements and design specifications. The developed system integrates various activities, including participant registration, program and class management, scheduling, assignments, quizzes, attendance tracking, assessment, payment processing, and information dissemination based on user access rights. Additionally, the system incorporates gamification features specifically Points, Badges, and Leaderboards (PBL) where Points are awarded for participant activities, Badges for specific achievements, and Leaderboards based on accumulated Points. The PBL mechanism is designed to provide recognition, feedback, and progress updates, thereby encouraging participants to interact more actively with the system. Black Box Testing demonstrated that the E-CRM's core functions and the PBL mechanism operate according to the designed scenarios. However, this study did not conduct a quantitative user evaluation to measure changes in user engagement. Consequently, the findings indicate that the E-CRM and PBL were successfully implemented to support service management and participant activities, although the effectiveness of PBL in measurably enhancing user engagement requires further evaluation.
Co-Authors Arif Wibowo Bambang Panca Syahputra Hambali Hambali Irianto Irianto Jeperson Hutahaean Lutfi Anniswa Sitorus Muhammad Amin Nasrun Marpaung Nurhaisyah Putri Oktaviya Siregar Rahmad Riyadi Raja Tama Andri Agus Rika Nofitri Riki Andri Yusda sahren, sahren Suparmadi Suparmadi Wiwin Handoko Zulkarnain Sirait Zulkhairani Zulkhairani