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
 43   44   45   46   47 
Brushstroke Classification from Oil Painting Images Using Convolutional Neural Networks for Tool Optimization Chantanasut, Suraphan
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.v6i4.1081

Abstract

This research introduces an integrated framework that applies Convolutional Neural Networks (CNNs) to classify brushstroke types from oil painting images and utilizes the classification results to inform the design and optimization of painting tools. Researchers will conduct the brush testing activities in four sessions: Session 1: Still life painting test, Session 2: Portrait painting test, Session 3: Landscape painting test, and Session 4: Rose painting test.The classification results were mapped to specific ergonomic and functional brush design parameters, resulting in the production of ten custom-designed brush prototypes. These brushes were fabricated using precision prototyping techniques and evaluated by twenty art students and five professional artists. Quantitative user feedback revealed high satisfaction across all performance categories, including ergonomic comfort, stroke control, and paint handling. The findings confirm that CNN-based analysis of brushstroke characteristics can directly support the practical innovation of art tools, bridging computational visual analysis and traditional artistic practice. This study offers a data-driven approach to creative tool design and presents a new interdisciplinary pathway that combines deep learning, material design, and fine arts.
Adaptive Neural Collaborative Filtering with Textual Review Integration for Enhanced User Experience in Digital Platforms Efrizoni, Lusiana; Ali, Edwar; Asnal, Hadi; Junadhi, Junadhi
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.v6i4.944

Abstract

This research proposes a hybrid rating prediction model that integrates Neural Collaborative Filtering (NCF), Long Short-Term Memory (LSTM), and semantic analysis through Natural Language Processing (NLP) to enhance recommendation accuracy. The main objective is to improve alignment between system predictions and actual user preferences by leveraging multi-source information from the Amazon Movies and TV dataset, which includes explicit user–item ratings and textual reviews. The core idea is to combine three complementary processing paths—(1) user–item interaction modeling via NCF, (2) temporal dynamics capture through LSTM, and (3) semantic understanding of reviews using NLP—into a unified deep learning-based adaptive architecture. Experimental evaluation demonstrates that this multi-input approach outperforms the baseline collaborative filtering model, with the Mean Absolute Error (MAE) reduced from 1.3201 to 1.2817 (a 2.91% improvement) and the Mean Squared Error (MSE) reduced from 2.2315 to 2.1894 (a 1.89% improvement). Training metrics visualization further shows a stable convergence pattern, with the MAE gap between training and validation consistently below 0.03, indicating minimal overfitting. The findings confirm that integrating cross-dimensional signals significantly enhances predictive performance and can contribute to increased user satisfaction and engagement in recommendation platforms. The novelty of this work lies in the simultaneous integration of interaction, temporal, and semantic dimensions into a single adaptive recommendation framework, a configuration not jointly explored in prior studies. Moreover, the flexible architecture enables adaptation to other domains such as e-commerce, music, or online learning, broadening its practical applicability.
Enhancing Autonomous Vehicle Navigation in Urban Traffic Using CNN-Based Deep Q-Networks Windarto, Agus Perdana; Solikhun, Solikhun; Wanto, Anjar
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.v6i4.896

Abstract

This research proposes a CNN-based Deep Q-Network (CNN-DQN) model to enhance the navigation capabilities of autonomous vehicles in complex urban environments. The model integrates CNN for spatial abstraction with reinforcement learning to enable end-to-end decision-making based on high-dimensional sensor data. The primary objective is to evaluate the impact of CNN-DQN state abstraction on the quality and stability of the resulting policy. Using a grid-based simulator, the agent is trained on a synthetic dataset representing urban traffic scenarios. The CNN-DQN model consistently outperforms standard DQN in multiple metrics: cumulative reward increased by 14.3%, loss convergence accelerated by 22%, and mean absolute error (MAE) reduced to 0.028. Furthermore, the model achieved a Pearson correlation coefficient of 0.94 in predicted actions and demonstrated superior robustness under Gaussian noise perturbation, with reward loss limited to 6.18% compared to 18.7% in the baseline. Visualizations of CNN feature maps reveal spatial attention patterns that support efficient path planning. The action symmetry index confirms that the CNN-DQN agent exhibits consistent left-right decision behavior, validating its policy regularity. The novelty of this study lies in its combined use of deep spatial encoding and value-based reinforcement learning for structured, rule-based environments with real-time control implications. These findings indicate that CNN-enhanced reinforcement learning architectures can significantly improve autonomous navigation performance and robustness in dynamic urban settings.
A Comparative Study of Machine Learning Approaches to Megathrust Earthquake Prediction in Subduction Zones Wella, Wella; Desanti, Ririn Ikana; Suryasari, Suryasari
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.v6i4.904

Abstract

Megathrust earthquakes are one of the most severe threats to countries situated along tectonic subduction zones, particularly Indonesia, where the movement of converging plates frequently triggers large-scale seismic events and tsunamis. Although recent developments in seismology have introduced various predictive tools, many of these models still face challenges, especially due to limitations in hydrogeological data quality. This study aims to investigate how three different machine learning algorithms perform in predicting megathrust earthquake events. The algorithms tested are Support Vector Machine, Random Forest, and Artificial Neural Network, applied to a dataset dominated by earthquake records from the Indonesian and Pacific regions. Each model was evaluated based on accuracy, precision, recall, and F1 score to provide a comprehensive performance analysis. The results show that Random Forest produced the highest accuracy, reaching 96%, followed closely by Support Vector Machine with 95%, while Artificial Neural Network achieved 83%. In terms of the F1 score, Random Forest led with a score of 0.95, indicating balanced performance in classification. However, recall, which is critical in disaster preparedness because it measures the model’s ability to detect high-risk events, Artificial Neural Network reached 92% for tsunami-related classifications. This suggests that while Random Forest is the most accurate overall, Artificial Neural Network could be more appropriate for early warning systems where the cost of missing a true event is much higher than issuing a false alarm. The contribution of this research is the direct comparison of multiple machine learning methods using real earthquake data, focusing not only on accuracy but also on practical disaster management considerations such as recall. This study also presents a novel perspective by analyzing the trade-off between model accuracy and disaster risk, emphasizing the need for probabilistic forecasts that can support timely public decision-making during seismic crises.
Analyzing the Efficacy of Pose Recognition, YOLOv3, and Deep Learning Techniques for Human Activity Recognition Zhumasheva, Ainur; Mansurova, Madina; Amirkhanova, Gulshat; Tyulepberdinova, Gulnur
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.v6i4.797

Abstract

The global increase in life expectancy, driven by increased nutrition, healthcare, and living conditions, has resulted in a significant growth in the senior population, notably in Kazakhstan, where the number of people aged 60 and more currently exceeds 2.7 million. This demographic transition poses considerable public health problems, particularly the high prevalence and severity of falls in older persons. Falls are currently the second largest cause of unintentional mortality for more than 87% of the elderly, with 28-34% falling at least once per year. As the worldwide population of people aged 65 and more is predicted to exceed 1.5 billion by 2050, there is an urgent need for precise, real-time fall detection systems. This work uses standardized datasets to conduct a complete evaluation of three fall detection methodologies: posture recognition, YOLOv3-based detection, and deep learning. Deep learning models attained the best accuracy of 92.0% by utilizing their capacity to learn complex spatial-temporal information, but at the cost of increased computing burden and slower inference times (40 ms). YOLOv3 provided competitive accuracy (90.2%) and quicker processing (25 ms), making it suitable for real-time deployment, although with a larger false positive rate. Pose identification, while highly interpretable due to its emphasis on skeletal key points, performed less well in crowded or obscured settings. The findings highlight the possibility for combining the capabilities of each technique to create hybrid systems with adaptive, resource-efficient architectures. Future research should focus on sensor fusion and optimization methodologies to improve accuracy and scalability across a variety of scenarios.
Mobile-Based AI Platform Integrating Image Analysis and Chatbot Technologies for Rice Variety and Weed Classification in Precision Agriculture Nuankaew, Wongpanya S.; Kuisonjai, Saweewan; Keawruangrit, Raksita; Nuankaew, Pratya
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.v6i4.918

Abstract

This work presents the development of an intelligent chatbot system capable of identifying rice plants and weeds from aerial photographs captured by smartphones, thereby enhancing precision agriculture. The study involves creating an AI model that utilizes image processing and deep learning techniques. Users can access the model through a LINE chatbot, and the study will also assess users' satisfaction with the model. Researchers gathered 12,000 pictures of rice fields in Phayao Province, Thailand, to train a modified InceptionV3 model using transfer learning. The dataset included images of rice plants and various types of weeds. The model was trained using image data collected under natural lighting and augmented to improve generalization. It achieved training, validation, and testing accuracies of 98.79%, 96.08%, and 97.83%, respectively. When deployed through a LINE Chatbot, it analyzed user-submitted images to estimate rice-to-weed ratios, yielding 73.33% average accuracy with consistent rice detection. Thirty individuals who used the system reported that it functioned well, was user-friendly, and provided significant benefits for farming in real-world applications. These results suggest that the system could leverage easily accessible AI tools to enhance farming efficiency, reduce costs, and positively impact the environment.
Application of Convolutional Neural Networks for Automated Iris Edge Detection in Sleepiness Monitoring during Blended Learning Tukino, Tukino; Yuhandri, Yuhandri; Sumijan, Sumijan
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.v6i4.882

Abstract

This study introduces a novel lightweight Convolutional Neural Network (CNN) model, T-Net, designed for real-time drowsiness detection based on eye closure patterns. The model was developed to address the prevalent issue of student fatigue in resource-constrained environments, such as during prolonged online learning or blended learning sessions. Unlike traditional deep learning models, T-Net prioritizes efficiency while maintaining high accuracy, making it suitable for deployment on devices with limited computational resources. The model uses a 68-point facial landmark detection technique to extract the eye region and accurately classify eyelid states (open or closed). Evaluated on two benchmark datasets, Dataset-1 (342 eye images) and Dataset-2 (1,510 eye images), T-Net demonstrated superior performance, achieving classification accuracies of 99.33% and 99.27%, respectively, outperforming other pre-trained models such as VGG19, ResNet50, and MobileNetV2. Usability testing revealed a high acceptance rate, with a System Usability Scale (SUS) score of 84.5, indicating the system’s practicality for real-world use. Additionally, statistical analysis showed a significant correlation (r = 0.67, p 0.01) between prolonged screen time and the emergence of visual fatigue symptoms. This study highlights the effectiveness of a lightweight CNN approach for real-time fatigue monitoring, offering a balance between performance and computational efficiency. The results suggest that T-Net can be effectively integrated into student monitoring systems to ensure alertness during learning sessions. Future research will focus on expanding the dataset, integrating infrared imaging for low-light environments, and incorporating additional fatigue indicators such as yawning and head pose.
Structural Equation Modeling Factors That Influence Online Course Acceptance in University Students Yuniansyah, Yuniansyah; Handayani, Febria Sri; Hartati, Eka; Aprizal, Yarza
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.v6i4.863

Abstract

Online courses are currently growing rapidly, influenced by various factors, especially among college students. This study uses the Technology Acceptance Model, by adding the variable of facilitating conditions. This study aims to determine the effects of perceived ease of use, perceived usefulness, and facilitating conditions on the actual use of online courses, mediated by intention to use. This study is the first conducted on students in university in South Sumatra, Indonesia, who have used or are currently using online courses. The data collection was carried out from June 24, 2024, to August 9, 2024. The research instrument used was a questionnaire based on a 5-point Likert scale. The sampling technique used in this study was purposive sampling, targeting students at universities under the Higher Education Service Institution Region II in Indonesia. A total of 360 students participated as respondents. The questionnaire results were analyzed using Structural Equation Modeling – Partial Least Squares (SEM-PLS). Data analysis was conducted using the SmartPLS software, version 4.1.1.2. The results showed that perceived ease of use had a significant influence on intention to use (path coefficient = 0.281, T-statistic = 6.642, P-value = 0.000), and perceived usefulness also had a positive influence on intention to use (path coefficient = 0.155, T-statistic = 4.545, P-value = 0.000). Facilitating conditions also had a positive influence on intention to use (path coefficient = 0.476, T-statistic = 8.880, P-value = 0.000), and intention to use significantly influenced actual use (path coefficient = 0.452, T-statistic = 10.490, P-value = 0.000). These findings highlight the important role of perceived ease of use, perceived usefulness, and facilitating conditions—mediated by intention to use—in significantly influencing the actual use of online courses among students, particularly at universities under the jurisdiction of the Higher Education Service Institution Region II in Indonesia.
Hybrid CNN Approach for Post-Disaster Building Damage Classification Using Satellite Imagery Sonang, Sahat; Yuhandri, Y; Tajuddin, Muhammad
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.v6i4.931

Abstract

Accurate post-disaster building damage assessment is critical for timely response and effective reconstruction planning. This study proposes a hybrid deep learning architecture that integrates Inception-ResNet-v2 and EfficientNetV2B0, designed to enhance post-disaster damage classification from high-resolution satellite imagery. The model leverages dual-stream feature extraction, followed by concatenated fully connected layers optimized with dropout and batch normalization to improve generalization and reduce overfitting. The objective is to outperform standard Convolutional Neural Network (CNN) models in terms of classification and segmentation performance across multiple damage categories: no damage, minor damage, major damage, destroyed, and unclassified. The model was trained and validated on the publicly available xView dataset, covering over 12,000 annotated images from various natural disasters. Comparative evaluation against ResNet, GoogleNet, DenseNet, and EfficientNet demonstrates that the proposed model achieves the highest accuracy (86%), precision (85%), recall (86%), and F1-score (84%). Furthermore, it outperforms all baseline models in segmentation metrics, achieving an Intersection over Union (IoU) score of 0.7749 and a Dice Similarity Coefficient (DSC) of 0.8726. The model also significantly reduces misclassification rates in critical categories such as “major damage” and “destroyed.” A Wilcoxon signed-rank test confirmed that these improvements are statistically significant (p 0.05) across all major performance indicators. The novelty of this study lies in the fusion of two state-of-the-art CNN backbones with tailored architectural modifications, yielding a robust and generalizable model suitable for automated disaster damage assessment. This research contributes a scalable deep learning approach that can be integrated into real-time or semi-automated disaster response systems, offering improved decision-making support in emergency contexts. The results affirm the model’s potential as a reliable tool in post-disaster scenarios and set a foundation for future work in multi-modal and real-time AI-based disaster management.
Study of Machine Learning Techniques for Predicting Panic Attacks with EEG and Personalized Binaural Beat Frequencies Batumalay, Malathy; Lakshmi Balaji, R S; Yingthawornsuk, Thaweesak
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.v6i4.759

Abstract

Panic attack detection and intervention remain critical challenges in mental health care due to their unpredictable nature and individual variability. This study proposes a machine learning-based framework for early detection of panic attacks using EEG-derived physiological signals, coupled with real-time personalized auditory intervention through binaural beat frequencies. Data were collected under controlled conditions using wearable biosensors to capture features such as heart rate variability, electrodermal activity, and skin temperature. A Gradient Boosting Classifier achieved 96% accuracy in detecting panic states, while an Isolation Forest algorithm effectively identified anomalous patterns preceding attacks. Based on physiological profiles, the system dynamically recommends individualized binaural beat frequencies to promote relaxation and emotional stabilization. The results demonstrate the feasibility of combining predictive modeling and neuroadaptive sound therapy to deliver scalable, non-invasive, and personalized mental health interventions. This approach aligns with global preventive health strategies, particularly those promoting digital therapeutics and early intervention for anxiety-related conditions.

Page 45 of 52 | Total Record : 518

 43   44   45   46   47 

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