cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Husni Teja Sukmana
Contact Email
husni@bright-journal.org
Phone
+62895422720524
Journal Mail Official
jads@bright-journal.org
Editorial Address
Gedung FST UIN Jakarta, Jl. Lkr. Kampus UIN, Cemp. Putih, Kec. Ciputat Tim., Kota Tangerang Selatan, Banten 15412
Location
Kota adm. jakarta pusat,
Dki jakarta
INDONESIA
Journal of Applied Data Sciences
Published by Bright Publisher
ISSN : -     EISSN : 27236471     DOI : doi.org/10.47738/jads
Core Subject : Science, Engineering,
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management
One of the current hot topics in science is data: how can datasets be used in scientific and scholarly research in a more reliable, citable and accountable way? Data is of paramount importance to scientific progress, yet most research data remains private. Enhancing the transparency of the processes applied to collect, treat and analyze data will help to render scientific research results reproducible and thus more accountable. The datasets itself should also be accessible to other researchers, so that research publications, dataset descriptions, and the actual datasets can be linked. The journal Data provides a forum to publish methodical papers on processes applied to data collection, treatment and analysis, as well as for data descriptors publishing descriptions of a linked dataset.
Arjuna Subject : Ilmu Komputer (semua kategori) - Kecerdasan Buatan
Articles 518 Documents
 47   48   49   50   51 
Software to predict maternal and child health risks with machine learning Fitriana, Fitriana; Zulkifli, Zulkifli; Rahayu, Sri
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1088

Abstract

Objective: Maternal healthcare services are essential in public health, prioritizing the health and well-being of women throughout pregnancy, childbirth, as well as the postpartum period. The services include various efforts to safeguard the health of both the mother and the unborn child. During these stages, mothers face numerous risks and complications, making early risk detection critical for ensuring the safety of the pregnancy. Method: A novel method is needed that enables more accurate and affordable screening to improve early detection as well as increase maternal and child healthcare. Therefore, this study aimed to propose a solution including the development of software that uses the Naive Bayes algorithm to predict maternal and child health risks. The perceptions provided by the application served as an initial diagnostic reference for both expectant mothers and healthcare providers, offering a cost-effective as well as precise alternative. Result: During the analysis, the Naive Bayes algorithm was compared with Neural Network (NN) and Random Forest (RF) models to evaluate the prediction accuracy. Among the models used, NN produced the lowest accuracy at 48%. Conclusion: The estimated cost for developing this application was IDR 1,635,913.
Image-Based Detection of Reduced Security Features in Indonesian Banknotes Using U-Net Architecture Andini, Silfia; Tukino, Tukino
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1087

Abstract

The circulation of fake currency banknotes in Indonesia continues to rise alongside rapid technological advancements, while conventional verification systems remain limited and often ineffective in detecting subtle authenticity cues. The main objective of this study is to develop an image-based fake currency detection system using the U-Net deep learning architecture and its modified version, T-Net, to enhance feature extraction and classification accuracy. The key contribution of this research lies in combining convolutional architectures with a practical, web-based interface that enables real-time image analysis, thus bridging the gap between model performance and user accessibility. A quantitative experimental method was employed, involving model development in Python using TensorFlow and Keras, and implementation of a Flask-based web application for real-time classification. The research utilized a dataset of 2,141 Indonesian rupiah banknote images, consisting of 1,015 genuine and 1,126 fake currency samples synthetically generated through digital modification of security features such as watermarks and color-shifting ink. Image preprocessing included resizing, normalization, and augmentation techniques such as random flipping and brightness adjustment to enhance data quality. Three convolutional architectures U-Net, ResNet-50, and the modified T-Net were trained and compared using identical hyperparameters. The T-Net model achieved the best performance, with 97.8% training accuracy, 82.6% validation accuracy, precision of 0.83, recall of 0.80, and an F1-score of 0.81. Despite the performance gap indicating overfitting, the model effectively distinguishes genuine from fake currency notes. The Flask-based interface allows users to upload images and receive classification results from all three models within 0.3–1.8 seconds per image. The findings demonstrate the feasibility and efficiency of U-Net based architectures for image-driven fake currency detection and provide a foundation for developing advanced, reliable, and real-time financial authentication systems that can strengthen digital security infrastructures in future applications.
EagleEyes: An Artificial Intelligence-Based Approach for Automatic Traffic Violation Detection Using Deep Learning Gata, Windu; Haris, Muhammad; Prasetiyowati, Maria Irmina; Harianto, Sony
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1097

Abstract

Rapid urbanization and the advancement of smart city programs in Indonesia necessitate intelligent, automated solutions for traffic monitoring and law enforcement. This study introduces EagleEyes, an artificial intelligence–based framework designed for automatic detection of multiple traffic violations by integrating the YOLOv8 deep learning architecture with Optical Character Recognition (OCR) for vehicle license plate identification. YOLOv8 was selected due to its anchor-free design, decoupled detection head, and enhanced feature fusion modules, which collectively improve detection accuracy, convergence speed, and small-object recognition compared to YOLOv5 and YOLOv7, while maintaining lightweight computational efficiency suitable for real-time applications. The proposed system was trained on a multi-class dataset representing common Indonesian violations, including seat belt non-compliance, helmet absence, motorcycle overcapacity, and unreadable license plates. Experimental results demonstrate robust performance, achieving a precision of 0.91, recall of 0.92, and mean average precision (mAP@0.5) of 0.96 at the optimal epoch, with an average inference speed of 25 frames per second and total training time of approximately 15 minutes on an NVIDIA RTX GPU. The OCR module attained an average recognition accuracy of 98.7%, although its performance decreased for vehicles captured beyond a five-meter distance due to reduced clarity and illumination inconsistencies. Implemented as a web-based application using the Flask framework, EagleEyes enables flexible browser-based visualization, and can be seamlessly integrated into Indonesia’s Electronic Traffic Law Enforcement (ETLE) infrastructure. Overall, the system demonstrates high potential to enhance smart city traffic management through scalable, AI-driven, and ethically responsible automation.
Applied Data Science for Exploring Multi-Channel Retail Service Quality Affecting Customer Satisfaction and Loyalty at Commercial Banks Le, Man Thi; Thanh, Tam Phan
Journal of Applied Data Sciences Vol 6, No 4: December 2025
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v6i4.1134

Abstract

This study examines how service quality across physical and digital channels influences customer satisfaction and loyalty within the omnichannel environment of commercial banks in Vietnam. Although digital transformation has accelerated rapidly, there remains limited empirical evidence on how integrating traditional service encounters with online and mobile platforms shapes customer perceptions and behavioral intentions. Addressing this gap, the paper develops and tests a comprehensive model that integrates traditional service quality dimensions, digital platform quality, and multi-channel integration, while also considering the moderating role of customers’ digital competence. The study contributes to the literature by extending conventional service quality frameworks to encompass the realities of omnichannel banking in an emerging market. It highlights the relative importance of physical facilities, staff professionalism, digital platform usability, and cross-channel consistency in shaping customer experiences. A two-phase methodology was employed. The qualitative phase involved expert evaluations and customer focus groups to refine measurement items and ensure contextual relevance. The quantitative phase gathered data from 785 retail banking customers and analyzed the relationships among the constructs using variance-based structural modeling. Findings indicate that all dimensions of service quality positively influence satisfaction, with physical facilities and multi-channel integration emerging as the strongest drivers. Satisfaction significantly enhances loyalty and mediates the effects of service quality dimensions. Digital competence both directly strengthens loyalty and moderates the satisfaction–loyalty relationship, suggesting that customers with higher digital skills derive more value from omnichannel services and are more likely to remain loyal. The study underscores the need for banks to invest in both modern physical infrastructures and high-performing digital platforms, while ensuring seamless integration across channels. It also emphasizes the importance of designing differentiated strategies tailored to customers’ digital capabilities to enhance overall satisfaction and foster long-term loyalty.
Assessing Large Language Models for Zero-Shot Dynamic Question Generation and Automated Leadership Competency Assessment Gheartha, I Gusti Bagus Yogiswara; Adiwijaya, Adiwijaya; Romadhony, Ade; Ardiansyah, Yusfi
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.970

Abstract

Automated interview systems powered by artificial intelligence often rely on fine-tuned models and annotated datasets, limiting their adaptability to new leadership competency frameworks. Large language models have shown potential for generating questions and assessing answers, yet their zero-shot performance, operating without task-specific retraining remains underexplored in leadership assessment. This study examines the zero-shot capability of two models, Qwen 32B and GPT-4o-mini, within a multi-turn self-interview framework. Both models dynamically generated questions, interpreted responses, and assigned scores across ten leadership competencies. Professionals representing the role of Digital Marketing and Account Manager participated, each completing two AI-led interview sessions. Model outputs were evaluated by certified experts using a structured rubric across three dimensions: quality of behavioral insights, relevance of follow-up questions, and fit of assigned scores. Results indicate that Qwen 32B generated richer insights than GPT-4o-mini (mean = 2.86 vs. 2.62; p less than 0.01) and provided more differentiated assessments across competencies. GPT-4o-mini produced more consistent follow-up questions but lacked depth in interpretation, often yielding generic outputs. Both models struggled with accurate scoring of candidate responses, reflected in low answer score ratings (Qwen mean = 2.35; GPT mean = 2.21). These findings suggest a trade-off between insight richness and scoring stability, with both models demonstrating limited ability to fully capture nuanced leadership behaviors. This study offers one of the first empirical benchmarks of zero-shot model performance in leadership interviews. It underscores both the promise and current limitations of deploying such systems for scalable assessment. Future research should explore competency-specific prompt strategies, fairness evaluation across demographic groups, and domain-adapted fine-tuning to improve accuracy, reliability, and ethical alignment in high-stakes recruitment contexts.
CNN-LSTM with Multi-Acoustic Features for Automatic Tajweed Mad Rule Classification Anggraini, Nenny; Rahman, Yusuf; Hidayanto, Achmad Nizar; Sukmana, Husni Teja
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1062

Abstract

The rules of mad recitation in the Qur’an are a crucial aspect of tajwīd, governing the lengthening of vowel sounds that affect both meaning and recitational accuracy. Despite its importance, there is currently no reliable automatic system capable of classifying mad rules based on voice input. This study proposes a deep learning-based approach using a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) model to automatically classify mad rules from Qur’anic recitations. The research follows the CRISP-DM methodology, covering data understanding, preparation, modeling, and evaluation stages. Acoustic features were extracted from 3,816 annotated audio segments of Surah Al-Fātiḥah, combining Mel-Frequency Cepstral Coefficients (MFCC), Chroma, Spectral Contrast, and Root Mean Square (RMS) to represent phonetic and prosodic attributes. The CNN layers captured spatial characteristics of the spectrum, while LSTM layers modeled temporal dependencies of the audio. Experimental results show that the combination of all four features achieved an accuracy of 97.21%, precision of 95.28%, recall of 95.22%, and F1-score of 95.25%. These findings indicate that multi-feature integration enhances model robustness and interpretability. The proposed CNN-LSTM framework demonstrates potential for practical deployment in voice-based tajwīd learning tools and contributes to the broader field of Qur’anic speech recognition by offering a systematic, ethically grounded, and data-driven approach to mad classification.
Self-consistency and Graph-based Filtering to Enhance Synthetic Arabic SMS Generation for Smishing Detection Alotaibi, Amal; Almasre, Miada; Surougi, Hadeel; Alkhozae, Mona; Alghanmi, Nouf
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1033

Abstract

Smishing or SMS phishing is a growing cybersecurity threat in mobile security, with Arabic-speaking regions particularly vulnerable due to the absence of large, labeled datasets. The main objective of this study is to develop a scalable pipeline that can generate and classify Arabic SMS messages to overcome the lack of data and enhance detection performance. The contributions are threefold: (i) constructing a balanced dataset of 6,903 messages by combining 903 synthetic samples with 6,000 real Arabic SMS messages; (ii) introducing a hybrid generation framework that integrates a fine-tuned GPT-3.5-turbo language model with Conditional WGAN embeddings, refined using self-consistency sampling and graph-based redundancy filtering; and (iii) evaluating the dataset using multiple machine learning (Logistic Regression, Random Forest, SVM) and deep learning (CNN, BERT) models. The pipeline unifies adversarial embedding generation, large language model fine-tuning, and cosine similarity filtering. Experimental results show consistently strong performance: Logistic Regression and Random Forest both achieved accuracy of 0.9949 and F1-score of 0.9950, while SVM outperformed all with accuracy 0.9957 and F1-score 0.9957. Among deep learning models, CNN reached accuracy 0.9942 and F1-score 0.9942, and BERT achieved 0.9900 across all metrics. These findings confirm that while SVM is most effective for this dataset, CNN and BERT add robustness by capturing semantic subtleties. Visual analyses, including confusion matrices and t-SNE projections, validated the overlap between real and synthetic embeddings, while comparative tables positioned this study within the context of recent Arabic smishing research. The novelty of this work lies in combining self-consistency and graph-based filtering within a hybrid generation-classification pipeline tailored for Arabic SMS, providing a reproducible framework extendable to low-resource, multilingual, and cross-platform environments such as WhatsApp and Telegram.
IndoBERT-SupCon: A Supervised Contrastive Learning Model for Analyzing Public Perception on Halal Tourism Octafia, Sri Mona; Malik, Rio Andika; Weriframayeni, Annisa; Delpa, Delpa
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1045

Abstract

The primary objective of this research is to develop and evaluate a robust deep learning model for accurately analyzing stakeholder perceptions of halal tourism development in Pariaman, West Sumatra, based on qualitative textual data. The main contribution is the introduction of IndoBERT-SupCon, a novel architecture that enhances the Indonesian BERT model with a Supervised Contrastive Learning (SupCon) mechanism. A novel method for producing more discriminative feature representations for complex viewpoints is presented in this paper, which is one of the first to use this sophisticated fine-tuning technique to Indonesian socio-political sentiment analysis. Conceptually, the model is trained to simultaneously minimize classification error while optimizing the feature space, pulling representations of similar sentiments closer together and pushing dissimilar ones further apart. To achieve this, we collected 1,022 primary textual responses through online surveys with tourists and in-depth interviews with key stakeholders, including SME owners and government officials. The SMOTE oversampling technique was employed on the training data to mitigate class imbalance. Experimental results on the test data demonstrate that the IndoBERT-SupCon model achieved outstanding performance, with a final accuracy of 96.59% and a macro F1-score of 0.97. These results significantly surpass the performance of a standard fine-tuned IndoBERT baseline, confirming the effectiveness of the SupCon approach. The findings provide the Pariaman local government with a highly valid, data-driven tool for more responsive and effective policy formulation. This research offers a robust framework that can be applied to other public policy domains, showcasing the value of advanced deep learning in transforming qualitative stakeholder feedback into actionable insights.
Improving University Ranking Robustness Using Rank Geometric Weight Integration with CoCoSo Method for Reducing Ordinal Weighting Instability Andryana, Septi; Mantoro, Teddy; Mutiara, Achmad Benny; Ernastuti, Ernastuti; Prihandoko, Prihandoko; Gunaryati, Aris
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1024

Abstract

This study lies in the field of decision support systems, focusing on the application of Multi-Criteria Decision Making (MCDM) for ranking alternatives based on predefined organizational criteria. A persistent challenge in this domain is the instability and subjectivity of ordinal weighting methods - such as Rank Order Centroid (ROC), Rank Sum (RS), Rank Reciprocal (RR), and Rank Order Distribution (ROD), which derive weights solely from rank positions, often leading to inconsistent and unreliable outcomes. To address this, this study introduces Rank Geometric (RG) weights, a geometric mean aggregation of ROC, RS, RR, and ROD designed to reduce subjectivity, stabilize weight distribution, and enhance robustness. By using the Combined Compromise Solution (CoCoSo) method, the RG against Times Higher Education’s (THE) official weights were evaluated, and the four individual ordinal methods, applied to the top 10 Indonesian universities across five THE 2025 ranking criteria. Empirical results show that RG-CoCoSo produces stronger and more consistent correlations with THE’s rankings than THE-CoCoSo, as validated by Spearman and Pearson correlation tests, with a p-value of 0.0251. This study contributes a practical, data-driven weighting framework that strengthens the reliability of MCDM-based institutional performance evaluation and can be generalized to other ranking contexts.
Multimodal AI Framework for Sign Language Recognition and Medical Informatics in Hearing-Impaired Patients Nuankaew, Pratya; Khamthep, Parin; Jaitem, Patdanai; Nuankaew, Kuljira S.; Nuankaew, Kaewpanya S.; Nuankaew, Wongpanya S.
Journal of Applied Data Sciences Vol 7, No 1: January 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i1.1096

Abstract

This study assesses the feasibility of YOLO-based detectors for the recognition of Thai Sign Language (TSL) within clinical intake workflows. We benchmark YOLOv5 through YOLOv10 over 100 to 150 training epochs and evaluate metrics including Precision, Recall, mAP@50, mAP@50:95, alongside training and validation losses to gauge stability. The losses decrease steadily as detection metrics improve; YOLOv10 offers the optimal balance, with Precision at 0.953, Recall at 0.939, mAP@50 at 0.933, and mAP@50:95 at 0.492. The improvements observed at stricter IoU thresholds are modest, underscoring ongoing challenges in achieving accurate localization and generalization across varying lighting conditions, viewpoints, occlusions, and motion. YOLOv11 has been excluded from the primary results due to abnormal loss behavior. These findings endorse a multimodal pipeline that employs an image-based detector as the central perception component, supplemented with pose and key point cues, as well as OCR and NLP layers, to transform recognized signs into structured medical intents for triage and telemedicine applications. Future research will focus on expanding sequence-level evaluation, incorporating dialects and co-articulation in TSL, and developing compressed or distilled models to facilitate reliable on-device inference in resource-constrained environments. 

Page 49 of 52 | Total Record : 518

 47   48   49   50   51 

Filter by Year

2020 2026


Reset
Filter By Issues
All Issue Vol 7, No 1: January 2026 Vol 6, No 4: December 2025 Vol 6, No 3: September 2025 Vol 6, No 2: MAY 2025 Vol 6, No 1: JANUARY 2025 Vol 5, No 4: DECEMBER 2024 Vol 5, No 3: SEPTEMBER 2024 Vol 5, No 2: MAY 2024 Vol 5, No 1: JANUARY 2024 Vol 4, No 4: DECEMBER 2023 Vol 4, No 3: SEPTEMBER 2023 Vol 4, No 2: MAY 2023 Vol 4, No 1: JANUARY 2023 Vol 3, No 4: DECEMBER 2022 Vol 3, No 3: SEPTEMBER 2022 Vol 3, No 2: MAY 2022 Vol 3, No 1: JANUARY 2022 Vol 2, No 4: DECEMBER 2021 Vol 2, No 3: SEPTEMBER 2021 Vol 2, No 2: MAY 2021 Vol 2, No 1: JANUARY 2021 Vol 1, No 2: DECEMBER 2020 Vol 1, No 1: SEPTEMBER 2020 More Issue