cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
-
Contact Email
-
Phone
-
Journal Mail Official
-
Editorial Address
-
Location
Kota pekanbaru,
Riau
INDONESIA
IT JOURNAL RESEARCH AND DEVELOPMENT
Published by Universitas Islam Riau
ISSN : 25284061     EISSN : 25284053     DOI : -
Core Subject : Science, Engineering,
Computer Science & IT Control & Systems Engineering Engineering
Information Technology Journal Research and Development (ITJRD) adalah Jurnal Ilmiah yang dibangun oleh Prodi Teknik Informatika, Universitas Islam Riau untuk memberikan sarana bagi para akademisi dan peneliti untuk mempublikasikan tulisan dan karya ilmiah di Bidang Teknologi Informatika. Adapun ruang lingkup dalam jurnal ini meliputi bidang penelitian di teknik informatika, ilmu komputer, jaringan komputer, sistem informasi, desain grafis, pengelolaan citra dan multimedia.
Arjuna Subject : -
Articles 6 Documents
 1 
Search results for , issue "Vol. 10 No. 2 (2025)" : 6 Documents clear
Comparative Analysis of SVM and XGBoost Classifiers with HOG Features for Concrete Crack Detection Adjie Eryadi, Ridha
IT Journal Research and Development Vol. 10 No. 2 (2025)
Publisher : UIR PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25299/itjrd.2025.20560

Abstract

This study offers a comparative assessment of the Support Vector Machine with Radial Basis Function Kernel and Extreme Gradient Boosting for automated concrete crack detection based on Histogram of Oriented Gradients feature extraction. Data comprised 40,000 RGB concrete images from an open-source Mendeley dataset; half were cracked and half were non-cracked. They processed through a preprocessing pipeline that includes the Poisson noise reduction and bilateral filtering techniques. Two approaches, holdout validation over several training/testing configurations (50:50, 60:40, 70:30, and 80:20) and systematic 5-fold cross-validation, were adopted for evaluation of the Wilcoxon signed-rank test for statistical significance and inference time for computational efficiency assessment. The experimental results indicate that SVM achieved a better holdout accuracy of 98.94% with the 80:20 configuration, while XGBoost achieved a cross-validation mean accuracy of 98.83% ± 0.0015. However, no statistically significant performance difference was revealed between the models according to the Wilcoxon analysis. Results indicated SVM excels at minimising false positives on undamaged surfaces, whereas XGBoost is better for identifying cracks, meaning that the choice of models used should depend on the application requirements, where applications require either the minimisation of false alarms or maximum sensitivity for detection in the case of structural health monitoring.
A Systematic Review on Machine Translation for Low Resource Nigerian Languages M. Abdulmusawir, Tijani; Kana, A. F. Donfack; Abubakar, Amina H.
IT Journal Research and Development Vol. 10 No. 2 (2025)
Publisher : UIR PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25299/itjrd.2025.21277

Abstract

Nigeria ranks among Africa's most linguistically diverse countries with over 500 indigenous languages, yet machine translation (MT) research remains severely limited for these low-resource languages. This systematic review examines the current state of MT research for Nigerian languages, identifies persistent challenges, and analyzes methodological trends. A systematic literature search was conducted across eleven databases including PubMed, Web of Science, and Scopus from January 2010 to August 2025. Search terms combined machine translation approaches with Nigerian language terms. Studies were screened using PRISMA guidelines requiring original research with evaluation metrics. From 51 papers, 25 duplicates were removed, 7 excluded for selection criteria, and 3 for lack of contribution, resulting in 16 studies. Only 11 Nigerian languages (2.2% of over 500 languages) were covered, creating a 97.8% research gap. Yoruba led with 4 studies, followed by Igala (3), Igbo and Nigerian Pidgin (2 each). Methods evolved from rule-based (4 studies, 2014 to 2021) through SMT (2 studies, 2016 to 2019) to NMT dominance (10 studies, 2018 to 2025). Idiomatic expression handling was the most persistent challenge (16.7%), followed by complex sentences, data scarcity, and domain specificity (each 9.5%). Nigerian MT research shows severe underrepresentation with persistent challenges in idiomatic expressions and data scarcity across all approaches. Neural method adoption reflects global trends but doesn't address resource constraints. Coordinated national approaches prioritizing parallel corpora creation and institutional partnerships are needed to prevent digital divides and support language preservation.
Comparative Evaluation of Machine Learning Models For Municipal Solid Waste Prediction With Feature Extension Milandile, Mwila; Sinyinda, Muwanei
IT Journal Research and Development Vol. 10 No. 2 (2025)
Publisher : UIR PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25299/itjrd.2025.22391

Abstract

This paper explores the utilization of machine learning approaches in predicting municipal solid waste generation accurately based on two distinct prediction methods, a single-model approach and a multi-model ensemble approach while incorporating feature engineering. Furthermore, we compare the predictive performance of two approaches: the single-model method and the multi-model ensemble approach. The metrics Mean Absolute Percentage Error (MAPE) the Root Mean Square Error (RMSE), and the Mean Absolute Error (MAE) have been used to assess the performance of the models. The finfings indicate that multi-ensemble approach outperformed the single model method by obtaining lower MAPE, RMSE, and MAE. The ensemble model obtained a Mean Absolute Percentage Error (MAPE) of 37.38 %, a Root Mean Square Error (RMSE) of 7610.76, and a Mean Absolute Error (MAE) of 5760.89, while the single-model technique achieved a Mean Absolute Percentage Error (MAPE) of 42.58 %, a Root Mean Square Error (RMSE) of 8258.01 and MAE of 6470.14. These findings indicate that merging multiple models can result in a more resilient and accurate predicting system. The findings presented in this paper suggest that by integrating feature engineering and utilizing multiple models results into more accurate predictions leading to effective waste management practices.    
Empirical Analysis of Deep Learning Models for Real-time Face Detection on Resource-constrained Devices Isong, Bassey; Ndouvhada, Sedzani; Kgote, Otshepeng
IT Journal Research and Development Vol. 10 No. 2 (2025)
Publisher : UIR PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25299/itjrd.2025.22402

Abstract

Face detection (FD) technology enables machines to identify human faces, playing a critical role in mobile device security and user interaction. However, achieving an optimal balance between speed and accuracy in FD algorithms remains a challenge, particularly for real-time applications on resource-limited devices. Factors such as variations in pose, lighting conditions, occlusions, dataset diversity, and hardware constraints often hinder effective deployment. This study presents a comprehensive empirical evaluation of deep learning-based object detection techniques, specifically YOLOv8, SSD, and Faster RCNN, to assess their effectiveness in addressing real-world scalability and performance demands. These models were trained on diverse datasets and evaluated using key performance metrics, including accuracy, precision, recall, and frames per second (FPS). YOLOv8 achieved superior performance, achieving 42.32 FPS with an accuracy of 86%, surpassing two-stage models in real-time processing speed while maintaining comparable accuracy. The findings underscore the importance of dataset quality and diversity in enhancing model performance and positioning YOLOv8 as an effective solution for balancing speed and accuracy on the COCO dataset. The study envisions a future exploration of hybrid models that integrate YOLOv8's efficiency with Faster RCNN's precision to develop more robust FD solutions tailored to real-world challenges.
Klasifikasi Daun Herbal Menggunakan Metode Convolutional Neural Network (CNN) Dengan Arsitektur VGG16 Auliani, Sisi; Indra, Dolly Indra; Fitriyani, Fitriyani Umar
IT Journal Research and Development Vol. 10 No. 2 (2025)
Publisher : UIR PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25299/itjrd.2025.24574

Abstract

Visual identification of various types of herbal leaves such as sweet leaf, moringa leaves, cat whiskers, bay leaf, and betel leaf is often difficult due to morphological similarities. This study presents a novel approach using a Convolutional Neural Network (CNN) model with the VGG16 architecture to automatically classify these five types of leaves. The main novelty of this study lies in the implementation of a two-stage fine-tuning strategy specifically tailored to the herbal leaf dataset. The first stage freezes the base layer and trains a new classification head, while the second stage fine-tunes several upper layers to improve model adaptability. The model was trained using 500 herbal leaf images and evaluated on 235 independent test images. The results demonstrated superior model performance with an overall accuracy rate of 91.06% and an average F1-score of 0.91. Qualitative analysis demonstrated the model's success in classifying leaves with unique features, such as cats whiskers and betel leaf. However, the model faced challenges in distinguishing leaves with high visual similarities, such as sweet leaf and moringa leaves. Practically, the developed model offers an effective and reliable solution for herbal leaf identification, reducing time and error rates compared to manual methods. Although this study is limited by the small dataset size, these results demonstrate the great potential of the optimized VGG16 architecture for applications in botany and traditional medicine, making it a valuable tool.
Enhancing Early Heart Disease Detection Through Comparative Analysis of Random Forest, Decision Tree, and K-NN Models Kohsasih, Kelvin Leonardi; Smith Sunario, Daniel; Alvin, Alvin; Laurendio, Fedro
IT Journal Research and Development Vol. 10 No. 2 (2025)
Publisher : UIR PRESS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25299/itjrd.2025.24703

Abstract

Heart disease is a leading cause of mortality worldwide and its rising prevalence challenges health systems. This study evaluates Decision Tree, k Nearest Neighbors, and Random Forest using the Heart Failure Prediction Dataset from Kaggle with 918 records and 12 demographic, clinical, and lifestyle features. The target variable indicates the presence of heart disease. Data preprocessing included cleaning, transformation, and scaling. Hyperparameters were tuned with stratified five fold cross validation to prevent data leakage. Performance was assessed using accuracy, precision, recall, F1 score, ROC AUC, PR AUC, Matthews Correlation Coefficient, and Brier score each estimated with 95 percent confidence intervals via bootstrap. k Nearest Neighbors achieved the highest accuracy at 90.2 percent, followed by Random Forest at 87.5 percent and Decision Tree at 85.3 percent. Calibration and decision curve analyses indicated that k Nearest Neighbors and Random Forest provided better calibrated probabilities and higher clinical utility across plausible thresholds. The study offers a reproducible evaluation pipeline and supports the use of machine learning for early detection of heart disease while encouraging future work on larger datasets and more advanced models.

Page 1 of 1 | Total Record : 6

 1 

Filter by Year

2025 2025


Reset
Filter By Issues
All Issue Vol. 10 No. 2 (2025) Vol. 9 No. 2 (2025) Vol. 9 No. 1 (2024) Vol. 8 No. 2 (2024) Vol. 8 No. 1 (2023) Vol. 7 No. 2 (2023) Vol. 7 No. 1 (2022) Vol. 6 No. 2 (2022) Vol. 6 No. 1 (2021) Vol. 5 No. 2 (2021) Vol. 5 No. 1 (2020) Vol. 4 No. 2 (2020) Vol. 4 No. 1 (2019) Vol. 3 No. 2 (2019) Vol 3 No 2 (2019) Vol 3 No 1 (2018) Vol. 3 No. 1 (2018) Vol 2 No 2 (2018) Vol. 2 No. 2 (2018) Vol 2 No 1 (2017) Vol. 2 No. 1 (2017) Vol 1 No 2 (2017) Vol. 1 No. 2 (2017) Vol. 1 No. 1 (2016) Vol 1 No 1 (2016) Article In Press More Issue