Article Per Year (5 Year)
p-Index From 2021 - 2026
0.751
P-Index
This Author published in this journals
All Journal JITK (Jurnal Ilmu Pengetahuan dan Komputer) Applied Information System and Management
Riki Winanjaya
Stikom Tunas Bangsa
Computer Science & IT Languange, Linguistic, Communication & Media

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

Found 3 Documents
Search

 1 
HYBRID TRANSFER LEARNING AND ADVANCED DATA AUGMENTATION FOR MULTICLASS BRAIN TUMOR CLASSIFICATION USING EFFICIENTNET A M H Pardede; Riki Winanjaya; Juni Ismail
JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer) Vol. 11 No. 3 (2026): JITK Issue February 2026
Publisher : LPPM Nusa Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/jitk.v11i3.7524

Abstract

Accurate Accurate brain tumor diagnosis from MRI images remains challenging due to dataset limitations, class imbalance, and high morphological variability across tumor types. Existing deep learning approaches often yield suboptimal results when trained on small or imbalanced datasets. This study proposes a hybrid learning strategy that integrates transfer learning with advanced data augmentation to classify four brain tumor categories: glioma, meningioma, pituitary adenoma, and normal tissue. Using a large-scale dataset of 7,023 MRI images, the proposed framework incorporates Mixup, CutMix, and a comprehensive augmentation pipeline with an optimized EfficientNet-B0 architecture. The model achieves a test accuracy of 99.05% with F1-scores of 0.99, representing a 4.05 percentage point improvement over a baseline InceptionV3 model (95.00%) and outperforming ResNet-based approaches (93.80%) reported in previous studies. This quantitative improvement demonstrates the effectiveness of combining modern CNN architectures with advanced augmentation strategies. The streamlined architecture and high accuracy make the method suitable for deployment in resource-constrained healthcare environments. These results indicate that hybrid augmentation and transfer learning can deliver clinically meaningful performance for early brain tumor identification, offering a scalable and practical solution for computer-aided medical diagnosis
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.
Application of EfficientNet Transfer Learning with Incremental Fine-Tuning for Road Damage Detection Riki Winanjaya; Abdi Rahim Damanik; Anton Abdulbasah Kamil
Applied Information System and Management (AISM) Vol. 9 No. 1 (2026): Applied Information System and Management (AISM)
Publisher : Depart. of Information Systems, FST, UIN Syarif Hidayatullah Jakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15408/aism.v9i1.50124

Abstract

Image-based road damage detection is an essential component of intelligent infrastructure monitoring systems. However, conventional transfer learning often fails to adapt pre-trained models to domain-specific characteristics such as fine-crack textures, illumination variations, and perspective distortions. This study proposes an EfficientNet-based road damage classification model that leverages incremental fine-tuning and multi-stage data augmentation to enhance feature adaptation and model robustness. The experiments were conducted using the Road Damage Detection dataset from Kaggle, comprising 1,400 labeled images across several road damage classes. The dataset was partitioned into 80:10:10 splits for training, validation, and testing, with stratification. The proposed approach gradually unfreezes EfficientNet layers through a structured incremental fine-tuning schedule while applying staged augmentation to expand data diversity. Experimental results show that the baseline EfficientNet transfer learning model achieved 78.26% accuracy, whereas the proposed model improved performance to 97.10% accuracy, with 97.60% macro precision, 97.20% macro recall, and 97.30% macro F1-score. The results demonstrate that incremental fine-tuning effectively enhances feature adaptation to road damage textures, while multi-stage augmentation improves model robustness. These findings indicate that the proposed approach provides an effective strategy for improving deep-learning-based road damage detection systems in real-world infrastructure monitoring applications.
Co-Authors Abdi Rahim Damanik Anton Abdulbasah Kamil Juni Ismail Pardede, Akim Manaor Hara Timbo Faritcan Parlaungan Siallagan Yusril Firas Alfaridzi Harahap