Article Per Year (5 Year)
p-Index From 2021 - 2026
3.441
P-Index
This Author published in this journals
All Journal JITK (Jurnal Ilmu Pengetahuan dan Komputer) Symmetry: Pasundan Journal of Research in Mathematics Learning and Education JOURNAL OF INFORMATION SYSTEM RESEARCH (JOSH) Didaktik : Jurnal Ilmiah PGSD STKIP Subang JOINT (Journal of Information Technology Jurnal Teknologi Informasi dan Komunikasi Bulletin of Computer Science Research International Journal of Management Science and Information Technology (IJMSIT) Abdi Cendekia: Jurnal Pengabdian Masyarakat Global Jurnal Manajemen Sistem Informasi Journal of Informatics, Electrical and Electronics Engineering Time in Physics: Journal for Theoretical, Instrumentation, Material-Molecular, and Education Physics Bulletin of Artificial Intelligence JuTISI (Jurnal Teknik Informatika dan Sistem Informasi)
Timbo Faritcan Parlaungan Siallagan
Universitas Mandiri
Religion Aerospace Engineering Agriculture, Biological Sciences & Forestry Arts Humanities Astronomy Automotive Engineering Biochemistry, Genetics & Molecular Biology Chemical Engineering, Chemistry & Bioengineering Chemistry Civil Engineering, Building, Construction & Architecture Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Dentistry Earth & Planetary Sciences Economics, Econometrics & Finance Education Electrical & Electronics Engineering Energy Engineering Environmental Science Health Professions Immunology & microbiology Industrial & Manufacturing Engineering Languange, Linguistic, Communication & Media Law, Crime, Criminology & Criminal Justice Library & Information Science Materials Science & Nanotechnology Mathematics Mechanical Engineering Medicine & Pharmacology Neuroscience Nursing Physics Public Health Social Sciences Transportation Veterinary Other

Published : 42 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 42 Documents
Search

 1   2   3   4   5 
Penerapan SMOTE pada Algoritma LightGBM dan XGBoost untuk Klasifikasi Penyakit Diabetes Regita Audyna Siregar; Solikhun Solikhun; Timbo Faritcan P. Siallagan
Journal of Informatics, Electrical and Electronics Engineering Vol. 5 No. 4 (2026): June 2026
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/jieee.v5i4.2593

Abstract

Diabetes mellitus is a chronic metabolic disease whose prevalence continues to increase worldwide and is often diagnosed late owing to insufficient recognition of its early symptoms. One of the major challenges in diabetes classification using machine learning is class imbalance in the dataset, which may cause the model to disproportionately favor the majority class and consequently compromise its capacity to accurately identify diabetic patients. Accordingly, this research seeks to compare the classification effectiveness of Light Gradient Boosting Machine (LightGBM) and Extreme Gradient Boosting (XGBoost) algorithms combined with the Artificial Minority Oversampling Approach (SMOTE) applied to diabetes classification task. The study utilized the Pima Indians Diabetes Dataset consisting of 768 instances and 8 attributes and followed the Machine Learning Life Cycle (MLLC), including exploratory data analysis, MinMax Scaler normalization, feature selection using Mutual Information, class imbalance handling with SMOTE, model training, and evaluation using accuracy, precision, recall, F1-score, and ROC-AUC as quantitative measures. Analysis of the results demonstrates that the application of SMOTE significantly improved the performance of both models. LightGBM with SMOTE achieved the best performance, with an accuracy of 85.33%, precision of 81.55%, recall of 91.33%, F1-score of 86.16%, and ROC-AUC of 90.49%, while XGBoost with SMOTE achieved an accuracy of 83.00%, precision of 79.64%, recall of 88.67%, F1-score of 83.91%, and ROC-AUC of 90.04%. These findings demonstrate that the combination of LightGBM and SMOTE is more effective than XGBoost in detecting diabetes cases and holds considerable promise for deployment as a clinical decision-making tool in early diabetes screening.
WEIGHTED LOSS STRATEGY FOR BERT-BASED TWITTER SENTIMENT ANALYSIS WITHOUT SYNTHETIC OVERSAMPLING Timbo Faritcan Siallagan; Riki Winanjaya; Juni Ismail
JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer) Vol. 12 No. 1 (2026): JITK Issue August 2026
Publisher : LPPM Nusa Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/jitk.v12i1.7680

Abstract

The widespread adoption of ChatGPT has generated extensive public discourse across social media, necessitating robust sentiment analysis to understand collective opinions. Traditional approaches frequently employ the Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance; however, its effectiveness on short-text data remains an open question. This study develops an optimized sentiment classification model and evaluates whether competitive performance can be achieved without synthetic data augmentation. The methodology encompasses comprehensive Natural Language Processing (NLP) preprocessing and stratified data partitioning to preserve distributional characteristics. A BERT-base architecture is fine-tuned using a class-weighted Cross-Entropy loss combined with weighted random sampling, deliberately avoiding SMOTE-based oversampling. The model is trained with the AdamW optimizer (learning rate: 3 × 10⁻⁵), batch size 32, and mixed-precision training for four epochs. On 198,639 preprocessed tweets, the proposed approach achieves 93.81% accuracy, with weighted precision, recall, and F1-score of 0.9365, 0.9381, and 0.9380 respectively, outperforming the baseline by 1.75 percentage points. Per-class analysis reveals strong performance for negative (F1-score: 0.96) and positive sentiment (F1-score: 0.94), with lower neutral classification (F1-score: 0.89), attributable to the inherent heterogeneity of neutral expressions. The training-validation gap remains below 5%, consistent with adequate regularization. These findings provide empirical evidence that, within the present experimental configuration, a properly optimized weighted loss strategy offers a viable and computationally efficient alternative to synthetic oversampling for BERT-based Twitter sentiment classification. Further controlled ablation studies and statistical validation are needed to establish generalizability.
Co-Authors Agnia Nurul Huda Agus Sudrajat Ahmad, Hermansyah Nur Alghifari, Muhammad Anderias Eko Wijaya Andrian, Avit Anestasya S, Sellyna Apandi, Jamaludin Asep Galih Miftah F. Dede Wisnu Dewi Untari Diah Lestari Dimas Juliana Tri N Ega Wahyu Andani Eka Susanti Endang Paza Pebrianti Eva Novia Triani Faelasivah, Fiky Fajar Fajar Ika Sriyanti Imin - Tugimin Iqbal Robiyana Jaja Jaja Juni Ismail Jupriyanto . Kodar Udoyono, Kodar Milawati, Fina Mirza Afif Pradivta Muhamad Agung Suhendra Muhamad Firmansyah Nita Wijaya Novianti, Wieke Nurul Utami Sudibyo Permana, Eka Rahma Dhea Safitri Regita Audyna Siregar Rian Hermawan Riki Winanjaya Sidik Abdul Miftah Siti Basrah Basrah SITI NURJANAH Solikhun Solikhun Solikhun Solikhun Suhendra, Muhamad Agung Tedi Sumardi Tita Tita Verdi Yasin Widiyanto Hermawan Wijaya, Handri Yanti Yanti