cover
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
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.
Articles 1,270 Documents
Mitigating Healthcare Information Overload: a Trust-aware Multi-Criteria Collaborative Filtering Model Qusai Y Shambour; Mosleh M Abualhaj; Ahmad Abu-Shareha; Abdelrahman H Hussein; Qasem M Kharma
Journal of Applied Data Sciences Vol 5, No 3: SEPTEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i3.297

Abstract

The rapid growth of online health information resources has made it difficult for users, as well as providers of healthcare, to cope with large volumes of information that are becoming increasingly complex. Hence, there is an urgent demand for developing new advanced recommendation techniques in the healthcare domain to enhance decision-making processes. However, most current health recommendation systems, which recommend personalized healthcare services and items such as diagnoses, medications, and doctors based on users' health conditions and needs, are hindered by the data sparsity issue that compromises the reliability of their recommendations. In this paper, we intend to address this issue by proposing a Trust-aware Multi-Criteria Collaborative Filtering model for recommendation services in the healthcare domain. This model leverages multi-criteria ratings and integrates user-item trust relationships to improve the precision and coverage of recommendations, thus facilitating more informed healthcare choices that align closely with their individual needs. Our empirical analysis on two healthcare multicriteria rating datasets, including those with sparse data, shows the proposed model's superior performance over existing baseline methods. On the RateMDs dataset, our model improved the average MAE by 24% and RMSE by 19% compared to baseline methods. For the WebMD dataset, it enhanced the average MAE by 6% and RMSE by 2%. In sparse data scenarios, the model boosted the average MAE by 18% and Coverage by 6% compared to baseline approaches.
Remote Photoplethysmography-Based HbA1c Estimation for Community Screening of Type 2 Diabetes Mellitus: Screening Tool or Diagnostic Alternative? Ernawati Ernawati; Yohanes Firmansyah; Alexander Halim Santoso; David Wongso; Diana Dinali; Muhammad Fikri Dzakwan; Bryan Anna Wijaya; Sri Tiatri; Putu Tommy Yudha Sumatera Suyasa; Meiske Yunithree Suparman
Journal of Applied Data Sciences Vol 7, No 4: December 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i4.1433

Abstract

Type 2 diabetes mellitus (T2DM) remains a major global health burden, with a high proportion of undiagnosed cases, particularly in community settings with limited access to laboratory testing. Remote photoplethysmography (rPPG) has emerged as a non-invasive, camera-based technology with potential for scalable screening. This study aimed to evaluate the agreement, diagnostic performance, and clinical role of rPPG-derived HbA1c as a screening tool or diagnostic alternative for T2DM. A cross-sectional diagnostic validation study was conducted among adults in a community setting in Semanan Subdistrict, Jakarta. Participants underwent non-contact facial scanning using an rPPG system alongside standard laboratory measurements of HbA1c and blood glucose. Agreement was assessed using Bland–Altman analysis and Cohen’s Kappa. Diagnostic performance was evaluated using sensitivity, specificity, predictive values, and receiver operating characteristic (ROC) curve analysis. rPPG-derived HbA1c showed a significant mean difference compared with laboratory measurements (0.63%; p 0.001; 95% CI: 0.34–0.93), indicating systematic bias. Categorical agreement was poor (κ = 0.029; p = 0.403). In screening analysis, sensitivity ranged from 65.38% to 97.37%, while specificity ranged from 8.91% to 59.55%, with consistently high negative predictive values (≥82%). ROC analysis demonstrated an AUC of 0.668 (p = 0.008) for diabetes and 0.711 (p 0.001) for prediabetes, indicating modest to acceptable discriminative ability. rPPG-based HbA1c estimation lacks sufficient agreement for diagnostic use. However, its relatively high sensitivity and scalability suggest that it may warrant further evaluation as a non-invasive screening tool for early detection and population-level risk stratification within mobile health settings.
Online Measuring Feature for Batik Size Prediction using Mobile Device: A Potential Application for a Novelty Technology Trianggoro Wiradinata; Theresia Ratih Dewi Saputri; Richard Evan Sutanto; Yosua Setyawan Soekamto
Journal of Applied Data Sciences Vol 4, No 3: SEPTEMBER 2023
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v4i3.121

Abstract

The garment industry, particularly the batik sector, has experienced significant growth in Indonesia, coinciding with a rise in the number of online consumers who purchase batik products through e-Marketplaces. The observed upward trend in growth has seemingly presented certain obstacles, particularly in relation to product alignment and product information dissemination. Typically, batik entrepreneurs originate from micro, small, and medium enterprises (MSMEs) that have not adhered to global norms. Consequently, the sizes of batik products offered for sale sometimes exhibit inconsistencies. The issue of size discrepancies poses challenges for online consumers seeking to purchase batik products through e-commerce platforms. An effective approach to address this issue involves employing a smartphone camera to anticipate body proportions, specifically the length and width of those engaged in online shopping. Subsequently, by the utilization of machine learning techniques, the optimal batik size can be determined. The UKURIN feature was created as a response to a specific requirement. However, it is essential to establish a methodology for investigating the elements that impact the intention to use this feature. This will enable developers to prioritize their feature development strategies more effectively. A total of 179 participants completed an online questionnaire, and subsequent analysis was conducted utilizing the Extended Technology Acceptance Model framework. The findings indicate that Perceived Usefulness emerged as the most influential factor. Consequently, when designing and developing the novelty feature of UKURIN, it is imperative for designers and application developers to prioritize the benefits aspect.
Spatio-Temporal Variation in Bus Service Satisfaction and Policy Implications: A Case-Based Study in an Emerging Urban Area Nguyen Thi Ngoc Nga; Hai Nguyen Minh
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.v6i4.879

Abstract

This study investigates the spatio-temporal variation in bus service satisfaction in a rapidly urbanizing city in Southeast Asia. The research addresses a critical gap in urban transit studies, where passenger satisfaction is often treated as a static construct, overlooking how satisfaction may fluctuate across different geographic areas and time periods. By applying a spatio-temporal analytical framework, this study aims to provide a more dynamic and localized understanding of transit service perceptions. The research builds upon existing approaches by integrating a structured survey based on the Customer Satisfaction Survey (CSS) method with Best–Worst Scaling (BWS) to prioritize service attributes. A stratified sampling technique was employed across multiple wards in the study area, with 612 valid responses collected during both peak and off-peak hours. The survey captured data on passenger experiences and preferences, disaggregated by time of day, location, and demographic characteristics. Multinomial Logit (MNL) modeling was used to estimate the relative importance of key service dimensions, such as punctuality, comfort, frequency, and accessibility. The analysis revealed significant spatial and temporal heterogeneity in satisfaction levels. For instance, passengers in peripheral wards rated reliability and onboard conditions more negatively compared to those in central areas. Similarly, satisfaction levels were lower during evening hours, particularly concerning bus overcrowding and wait times. The findings suggest that transit policy must adopt a more flexible and localized strategy, rather than uniform service standards, to address distinct user expectations. Targeted improvements in underperforming routes and time slots could enhance overall user experience and promote public transport usage. This study contributes new insights to the evaluation of urban bus services in emerging cities and underscores the value of incorporating spatio-temporal dynamics into transit planning and customer satisfaction research.
Optimizing Emergency Logistics Identification: Utilizing A Deep Learning Model in the Big Data Era V. Sumathi; B. L. Shivakumar; Siti Sarah Maidin; Wu Ge
Journal of Applied Data Sciences Vol 5, No 3: SEPTEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i3.369

Abstract

This study investigates the dynamics of commodity flow across different facilities and settings, evaluating the performance of Simulation and Feed Forward Neural Network (FFNN) methods in optimizing these flows. Analyzing data from various configurations, the research reveals significant variations in commodity distribution patterns. At Facility_1 from the K1 disposer market, the flow of Commodity_1 increased from 770 units to 830 units, while Commodity_2 decreased from 192 units to 166 units. At Facility_2, Commodity_1's flow decreased from 851 units to 793 units, and Commodity_2's flow slightly increased from 139 units to 148 units. Similar trends are observed at facilities from the K2 disposer market, reflecting the complex impact of different settings on commodity flow. The comparative analysis of Simulation and FFNN methods demonstrates their relative strengths. In Setting I, the Simulation method achieved an objective value of 1,800,574.36 Rs with a computation time of 46.78 seconds, while the FFNN method yielded a slightly lower objective value of 1,800,352.24 Rs in 42.01 seconds. In Setting II, the Simulation method provided an objective value of 1,801,025.36 Rs with a computation time of 103.86 seconds, whereas FFNN achieved a comparable objective value of 1,800,847.27 Rs in 63.05 seconds. In Setting III, Simulation resulted in an objective value of 1,801,527.36 Rs with a computation time of 61.12 seconds, while FFNN produced a higher objective value of 1,806,997.32 Rs in 50.03 seconds. The results highlight the trade-offs between solution quality and computational efficiency. The Simulation method consistently delivers higher objective values but requires more time, making it suitable for applications where result accuracy is crucial. Conversely, the FFNN method offers faster computation with competitive or superior objective values, making it advantageous for scenarios where time constraints are significant. This study underscores the importance of selecting appropriate computational methods based on specific operational needs, optimizing both the efficiency and effectiveness of commodity flow management.
Artificial Intelligence Applications and Digital Finance Development: The Moderating Role of Human Resources and Digital Infrastructure Huynh Cao Tuan; Tran Thanh Phuong
Journal of Applied Data Sciences Vol 7, No 2: May 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i2.1247

Abstract

The rapid advancement of digital technologies has fundamentally transformed financial systems, particularly in emerging economies, where digital finance plays a critical role in enhancing financial accessibility and efficiency. Among these technologies, artificial intelligence (AI) has emerged as a strategic driver reshaping digital financial services. This study aims to investigate the direct and moderating effects of artificial intelligence applications on the development of digital finance by integrating technological, human, institutional, and innovation perspectives. A sequential mixed-methods design was employed. In the qualitative phase, semi-structured interviews were conducted with 55 experts in banking, finance, and financial technology. The sample size was determined based on theoretical saturation, which was reached when successive interviews yielded no substantially new insights regarding construct dimensions or measurement refinement. Insights from this phase were used to validate constructs and refine the measurement instrument. In the quantitative phase, survey data were collected from 700 digital banking users across 20 commercial banks in Vietnam. Partial Least Squares Structural Equation Modeling (PLS-SEM) was applied to test the proposed hypotheses. The results indicate that the digital policy framework (β = 0.300) and the digital human resource management (β = 0.279) exhibit the largest direct effects on digital finance development, based on the relative magnitude of standardized path coefficients. Among these factors, digital policy frameworks and digital human resource management demonstrate the strongest direct impacts. More importantly, the findings confirm the moderating role of artificial intelligence applications. AI significantly strengthens the relationships between digital human resource management and digital finance development, as well as between digital technology infrastructure and digital finance development. These results indicate that AI serves not only as an independent technological driver but also as a strategic catalyst, enhancing the effectiveness of digital infrastructure and human capital. This study contributes to the digital finance and information systems literature by empirically demonstrating that artificial intelligence serves as both a determinant and a moderator of digital finance development. From a practical perspective, the findings suggest that policymakers and banking executives should prioritize AI-enabled human resource strategies and the intelligent use of digital infrastructure to accelerate the development of sustainable digital finance in emerging economies.
HU Variance Moment Optimizes Keyframe Selection Based on Deep Learning for Violence Detection Sukmawati Anggraeni Putri; Pulung Nurtantio Andono; Purwanto Purwanto; Moch Arief Soeleman
Journal of Applied Data Sciences Vol 6, No 2: MAY 2025
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v6i2.648

Abstract

Violence in public spaces poses a serious threat to individuals and society. Manual monitoring and violence detection require much time and human resources, ultimately hindering detection accuracy and speed. Therefore, an automated method is needed to detect violence to ensure fast and efficient action. Along with technological advances, violence detection research has adopted various methods and models, including deep learning, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs). In this study, the classification process for detecting violence and non-violence uses the VGG19 model, one of the CNN models that has good performance with limited computing. In addition, the Long Short-Term Memory (LSTM) model is the best RNN model for processing temporal data in videos. However, this performance will decrease with noise and irrelevant data in the classification process. Therefore, to optimize deep learning performance, this study in the pre-processing phase selects keyframes in frame extraction using the Hu Variance Moment Technique. This method calculates each frame’s Hu and Variance Moment values and selects keyframes based on high Hu values. Next, we use Adaptive Moment Estimation (Adam) to optimize the gradient of the selected keyframes. This study produces a Hu19LSTM model tested on three datasets: hockey fight, crowd, and AIRTLab. The proposed Hu19LSTM model produces an accuracy of 97% on the Hockey Fight dataset, 97% on the Crowd dataset, and 95% on the AIRTLab dataset. These results indicate that the Hu19LSTM model can increase its accuracy on the hockey fight and Crowd dataset by 97%.
Developing the Readiness and Success Model of Information System Implementation in the Indonesian Equestrian Industry Ajang Sopandi; Nor Adnan Yahaya; Aang Subiyakto
Journal of Applied Data Sciences Vol 5, No 1: JANUARY 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i1.145

Abstract

This study reports on the incorporation of technology readiness models in information system (IS) success models in the context of assessing readiness factors and the success of information system integration in the equestrian sports industry in Indonesia. As found in several information systems studies, many IS models are developed by adopting, combining, and adapting previous models. Researchers developed this model based on input-process-output logic and processional and causal models of information system success models. The developed model is structured by involving 12 variables and 62 indicators. The path of influence between variables is described by 30 links. In the research implementation stage, the model is also broken down into more detailed assessment instruments. Although these model development studies may have limitations on the assumptions used and the researchers' understanding, they can make theoretical contributions, particularly in terms of proposed new models. In addition, transparency in model development, proposed models, and data collection instruments may also be a practical consideration for advanced research in the context of readiness and successful implementation of information systems in the equestrian sports industry in Indonesia
An Explainable Credit Card Fraud Detection Model using Machine Learning and Deep Learning Approaches Mona Alkhozae; Miada Almasre; Abeer Almakky; Reemah M. Alhebshi; Amani Alamri; Widad Hakami; Lamia Alshahrani
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.v6i4.962

Abstract

This study proposes an adaptive, interpretable real-time fraud detection and prevention system designed for high-risk financial environments, capable of processing over 1.6 million imbalanced credit card transactions with low latency. The objective is to build a unified framework that integrates predictive accuracy, explainability, and adaptability. The methodology follows four phases: exploratory data analysis to reveal structural and behavioral fraud patterns, feature engineering with domain-informed attributes and ADASYN oversampling to mitigate the 1:174 imbalance, training of multiple models (XGBoost, LightGBM, Random Forest, Gradient Boosting, and MLP), and an ensemble architecture evaluated with SHAP-based explainability. The system introduces three key contributions: stability-aware SHAP caching that reduces explanation latency to 41.2 ms, reinforcement learning–based threshold tuning that dynamically adapts to evolving fraud patterns, and out-of-distribution detection to enhance resilience against data drift. Results demonstrate strong performance, with XGBoost achieving 99.86% accuracy, 96.36% precision, 80.59% recall, F1-score of 0.878, and ROC-AUC of 0.9988, outperforming other models. The full system attained 93.2% accuracy, 90.2% F1-score, and 96.1% AUC at the system level, successfully blocking 91% of fraudulent transactions while maintaining a false positive rate of 7.8%. Novelty lies in combining explainability and adaptivity in a production-ready architecture, where reinforcement learning enables continuous threshold self-regulation and SHAP stability analysis validates interpretability across models. These findings show that high fraud detection accuracy and transparency are not mutually exclusive, offering a scalable blueprint for financial institutions and other critical domains requiring real-time, explainable, and adaptive decision-making.
AI Prediction Model to Investigate the GovTech Maturity Index (GTMI) Indicators for Assessing Governments’ Readiness for Digital Transformation Yahya AlMurtadha
Journal of Applied Data Sciences Vol 5, No 4: DECEMBER 2024
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v5i4.373

Abstract

Digital transformation helps governments improve their efforts to provide services to beneficiaries around the clock. However, governments must consider the potential disadvantages of unplanned digital transformation such as lack of attention to cybersecurity standards, which could put citizens' data at risk, or resistance to change and adoption of new technologies by government employees. The goal is for governments to take a comprehensive, well-planned approach to digital transformation that addresses people, processes and technology. Hence, governments should utilize digital maturity models to assess their current state and develop a plan for successful digital transformation. Governments especially are seeking for a smart transition to a mature digital transformed state. Therefore, this study proposes using the digital transformation maturity index as a systematic framework for governments to assess their digital transformation and plan a successful digital transformation. This study suggests using AI prediction algorithms to chart a path for a mature digital transformation.  Hence, this study builds a model that predicts government maturity level to one of four maturity classes (A, B, C, and D) using several AI prediction techniques on the World Bank GovTech dataset, which contains 48 important indicators used to measure the GovTech maturity index.  The results show that decision tree algorithm outperforms other approaches in terms of prediction accuracy. Government’s experts may thus utilize decision trees to determine the digital transformation maturity index success route starting at the root and working their way up to the leaf.  The results also highlight the need for a government to examine three essential indicators for a successful digital transformation with higher maturity class: universally accessible citizen-centric public services, a national strategy to connect all departments under one goal, and transparency. The study concludes that governments should embrace holistic and well-planned digital transformation while considering factors such as cultural and behavioral changes, future disruptions and emerging technologies.