Article Per Year (5 Year)
p-Index From 2021 - 2026
0.61
P-Index
This Author published in this journals
All Journal INTI Nusa Mandiri Journal of Informatics Development
Abdur Ro’uf
Institut Teknologi dan Bisnis Widya Gama Lumajang
Computer Science & IT

Published : 3 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 3 Documents
Search

 1 
REKOMENDASI PEKERJAAN BIDANG EKONOMI : SISTEM REKOMENDASI MENGGUNAKAN CONTENT BASED Rouf, Abdur; Asy’ari , Hasyim; Yafi Urrohman, Maysas; Devi Rahmawati , Febriane
INTI Nusa Mandiri Vol. 20 No. 1 (2025): INTI Periode Agustus 2025
Publisher : Lembaga Penelitian dan Pengabdian Pada Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/inti.v20i1.6786

Abstract

The recommendation system was developed to assist students of the Institut Teknologi dan Bisnis Widya Gama Lumajang, particularly those from the Faculty of Economics and Business, in determining their preferred career options. This system helps students by providing various job references that match their individual criteria. The data was collected from a tracer study, which includes information such as academic grades, non-academic achievements, job positions, company names, salaries received. From the total dataset, 1,120 records were deemed valid and used in the research process. The aim of this research is to assist students by providing job recommendations based on similar criteria between current students and alumni. The method applied in this study is quantitative experimental research based on data mining, with the main approach being Content-Based filtering and the MLP (Multi-Layer Perceptron) Classifier algorithm. The data was split into two parts: 65% for training and 35% for testing. This division aims to allow the model to learn from most of the data while also being tested for accuracy using unfamiliar data. The recommendation model was developed using the MLP Classifier algorithm with a hidden_layer_size configuration of 100 neurons and a max_iter of 200 iterations. For the initial test, 10 sample data points were used to evaluate the model’s performance. During training, the loss value was monitored to assess how well the model understood the data and adjusted its internal weights. With this configuration, the system is expected to provide accurate job recommendations based on the user’s profile and academic history.
Cluster Analysis of Alumni Job Waiting Times Using the K-Means Algorithm with Elbow and Silhouette Methods Abdur Ro’uf; Hasyim Asy’ari; Maysas Yafi Urrochman
Journal of Informatics Development Vol. 5 No. 1 (2026): Oktober 2026
Publisher : Institut Teknologi dan Bisnis Widya Gama Lumajang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30741/jid.v5i1.2080

Abstract

Evaluating graduate employability is a key indicator for assessing the quality of higher education institutions. One effective approach involves analyzing alumni tracer study data using data mining techniques. This study aims to categorize alumni data based on the waiting period for employment after graduation using the K-Means algorithm. To determine the optimal number of clusters, two methods were employed: the Elbow method (Sum of Squared Errors/SSE) and the Silhouette Score. The research process included data preprocessing—comprising data selection, transformation, and normalization—to enhance data quality prior to clustering. Test results indicated that the Elbow method yielded an optimal cluster count of k=4, whereas the Silhouette method showed the best result at k=3, with a score of 0.62. Consequently, k=3 was selected as the optimal number of clusters due to superior cluster quality. The clustering results revealed that the majority of alumni (67.6%) experienced a short waiting period (averaging ±2 months), while 16.5% faced a long waiting period (±8 months), and 15.9% fell into the medium category (±5 months). These findings are expected to serve as a basis for higher education institutions to evaluate and improve graduate quality, ensuring better alignment with the demands of the job market.
Comparative Analysis of Transformer Models for Emotion Classification in Indonesian Text Data Faishal Basbeth; Cahyasari Kartika Murni; Abdur Ro’uf
Journal of Informatics Development Vol. 5 No. 1 (2026): Oktober 2026
Publisher : Institut Teknologi dan Bisnis Widya Gama Lumajang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30741/jid.v5i1.2082

Abstract

This study presents a comparative analysis of three transformer-based models, BERT, RoBERTa, and DistilBERT, for emotion classification on Indonesian text. The Indo4B dataset, which consists of five emotion labels (anger, fear, happiness, love, and sadness), is used as the benchmark corpus. All three models were trained using fine-tuning and hyperparameter-tuning procedures. The evaluation was conducted using accuracy, F1-score, model size, and training time as the primary performance metrics. The results indicate that RoBERTa achieves the best overall performance, with an accuracy of 90.83% and an F1-score of 91. Meanwhile, DistilBERT demonstrates approximately 60% faster training time compared to BERT, with only a 0.61% decrease in accuracy. These findings suggest that DistilBERT offers a highly efficient alternative while maintaining competitive classification performance.
Co-Authors Asy’ari , Hasyim Cahyasari Kartika Murni Devi Rahmawati , Febriane Faishal Basbeth Hasyim Asy’ari Maysas Yafi Urrochman Yafi Urrohman, Maysas