International Journal of Advances in Data and Information System
Not a member yet
161 research outputs found
Sort by
Enterprise Architecture Design Vision Architecture and Business Architecture Stage Using TOGAF ADM at SMA ABC
Full text linkDigitalization of Education in realizing the vision of an independent curriculum, encourages the development of educational unit technology. In realizing the development of information technology, enterprise architecture planning (EAP) is used to build a balance between business strategy and information technology in implementing integrated educational unit digitalization. This research was conducted at SMA ABC Bandung. The research method used is The Open Group Framework (TOGAF) Architecture Development Method (ADM) with the aim of explaining and providing a methodology in analyzing EAP and finding the vision architecture and business architecture periodically. The purpose of this study is to design information system management, direct business strategy with planning, information technology governance and educational unit management. Research results In the architectural vision, the research is able to identify the needs of organizational capacity that must be met to support the achievement of the vision and mission of SMA ABC Bandung. While at the Business Architecture stage, it produces a gap analysis (GAP analysis) of existing business processes, produces business process diagrams and functional decomposition diagrams to describe the main and supporting business processes at the school, produces stakeholder identification, organizational scope, and main business processes at SMA ABC, and formulates a catalog of principles that are the basis for the proposed enterprise architecture design
Comparison of Transfer Learning Using VGG16, MobileNetV2, and ResNet50 for Pornography Image Detection
Full text linkThe rapid growth of digital technology is vital for the Indonesian Scout to reach and interact with its members. The National Indonesian Scout (Kwarnas) uses the “Ayo Pramuka” social media application to support this. However, such platforms risk exposing users, especially teenagers, to harmful content like pornography. This research applies Computer Vision and Transfer Learning Convolutional Neural Networks (CNNs) to detect pornographic images automatically. The objective is to identify the CNN model (VGG16, MobileNet V2, ResNet 50) with the highest detection accuracy and determine the impact of color space preprocessing. The method includes two stages first, image preprocessing by converting RGB images to HSV and YCbCr second, feature extraction using pre-trained CNNs with freezing and fine-tuning. A dataset of 4060 images was used for training and testing. Without preprocessing, VGG16 achieved the best accuracy of 99.01%. When RGB images were converted to HSV, ResNet 50 produced the highest accuracy of 99.51%. The findings show that combining color space transformation and Transfer Learning CNN significantly improves pornographic content detection in the “Ayo Pramuka” Application, enhancing safe digital engagement for Indonesian Scouts
Ambidextrous AI Governance for Driving SmartCo’s Digital Transformation Using COBIT 2019 Traditional and DevOps
Full text linkArtificial Intelligence (AI) plays a vital role in accelerating digital transformation within the technology sector. This study investigates SmartCo, a technology company seeking to enhance the security and governance of AI implementation using an ambidextrous COBIT 2019 framework that integrates people, processes, and technology. The research adopts a Design Science Research (DSR) methodology, utilizing interviews, questionnaires, and internal document analysis until data saturation was achieved. Governance and Management Objectives (GMOs) were prioritized using design factors, DevOps practices, relevant regulations (ICT Minister Regulation No. 5/2021 and SOE Minister Regulation No. PER-2/MBU/03/2023), and previous studies. DSS05 (Managed Security Services) was selected as the primary focus, reflecting the organization’s priority on data protection and secure AI operations. The capability maturity assessment revealed gaps in security leadership, documentation, and process automation, indicating the need for more adaptive and integrated governance. Targeted improvements were implemented, including formalizing governance structures, enhancing security training, and adopting supportive technologies, which increased the DSS05 maturity level from 3.00 to 3.86. A comprehensive roadmap guides further enhancements in security-focused governance. This study provides practical insights for organizations aiming for secure, AI-enabled digital transformation. In addition, it contributes to the theoretical foundation of ambidextrous COBIT 2019 governance frameworks by demonstrating their application in a regulated technology environment
XGBoost Model Optimization Using PCA for Classification of Cyber Attacks on The Internet of Things
No full textThe rapid expansion of the Internet of Things (IoT) ecosystem has increased its susceptibility to cyberattacks, creating a critical need for reliable Intrusion Detection Systems (IDS). However, IDS performance is often hindered by severe class imbalance, high-dimensional features, and similarities among attack behaviors. This study proposes an optimized XGBoost model enhanced with the Synthetic Minority Over-sampling Technique (SMOTE) and Principal Component Analysis (PCA) to address these challenges. A systematic grid-search procedure was employed to ensure transparency, reproducibility, and optimal hyperparameter selection. The original imbalance ratio of approximately 1:27 was successfully normalized to nearly 1:1 through SMOTE. The Gotham dataset used in this study consists of roughly 350,000 IoT traffic records across eight attack categories. Five data-splitting scenarios (50:50 to 90:10) were evaluated using stratified hold-out validation supported by k-fold cross-validation. The optimized model achieved 99.68% accuracy, while extremely high AUC values approaching 1.0 were carefully validated to eliminate potential data leakage. Naive Bayes, Logistic Regression, Support Vector Machine, and Deep Neural Network were included as baseline comparisons. The results demonstrate that combining SMOTE and PCA significantly improves model stability and generalization on imbalanced IoT traffic, confirming the effectiveness of the proposed XGBSP method
A Content-Based Filtering Approach for Matching Village Potentials with Community Service Programs
Full text linkVillage development through the development of village potential is carried out by universities in the form of community service. With the community service program, it is expected to help villages in village development so that they can become independent villages. However, in its implementation, the designed assistance programs are not specific and not aligned with the needs and potential of the village. As a result, the assistance provided is less effective and having minimal impact on village development. One of the causes is the unavailability of data on village potential and problems systematically and structured. Based on these problems, a recommendation system is needed that is able to provide assistance program proposals that are in accordance with the potential and problems of the village specifically and relevantly. This research uses Content-Based Filtering which provides recommendations based on the similarity of input data content with available historical data. The purpose of this research is to make the planning and implementation process of village assistance programs more efficient, effective, and on target. The results of the research are that the Content-Based Filtering method has proven effective in providing recommendations that are appropriate for mapping village potential and village assistance programs
Analysis of Factors Influencing the Intention to Use QRIS As a Payment Tool in Central Kalimantan Province
Full text linkIn recent years, the Quick Response Code Indonesian Standard (QRIS) has emerged as a digital payment system in Indonesia. Central Kalimantan has witnessed a substantial increase of 91% in the number of merchants adopting this system in 2021 compared to previous years. However, the significant growth of merchants with QRIS options has not yet had the expected increase in usage frequency. This research aims to identify the factors influencing the utilization of QRIS as a payment method and to formulate recommendations to enhance its adoption. Employing a quantitative method, this study modifies the Technology Acceptance Model (TAM) by including external variables such as Subjective Norm, Perceived Security based on preliminary research to understand their impact on Behavioral Intention to use. Data calculation and analysis were conducted using the SmartPLS 3 tool. The findings reveal that Attitude Toward Using exerts a significant influence on Behavioral Intention to Use. While Subjective Norm, Perceived ease of use and Perceived Usefulness significantly affect usage intention through Attitude Toward Using, this study highlights the potential for increased QRIS adoption by leveraging community figures or influencers in socialization efforts. Furthermore, enhancing perceived usefulness through merchant promotions and user education is crucial for fostering positive attitudes towards QRIS usage
A Hybrid Model of Graph Attention Networks and Random Forests for Link Prediction in Co-Authorship Networks
Full text linkCo-authorship prediction is important in academic network analysis due to it helps to understand patterns of scientific collaboration and supports collaboration recommendation systems. Topology-based approaches, such as connectivity metrics and node distance, have been widely used to model new relationships in networks. However, these approaches often overlook relevant author attributes, such as reputation and productivity. This study develops a co-authorship prediction model by combining a Graph Attention Network (GAT) and a Random Forest. GAT is used to extract topological features from the co-authorship graph, while Random Forest leverages additional attributes such as h-index and the number of publications to improve prediction accuracy. Experiments were conducted on a co-authorship dataset comprising over 10,000 authors and 50,000 publications. The results show that GAT achieved 85% accuracy, while Random Forest reached 80%. The combination of the two yielded 90% accuracy and a higher F1-score, indicating a better balance between precision and recall. The combined model also proved more accurate in predicting collaborations involving highly productive authors. These findings suggest that a hybrid approach can more comprehensively capture the dynamics of academic collaboration and may serve as a foundation for developing more effective collaboration prediction systems in the future
Analysis of User Experience Usage on the Sardjito Hospital and Yogyakarta Regional Public Hospital Websites Using the User Experience Questionnaire (UEQ)
Full text linkThe user experience of healthcare websites is crucial for ensuring accessibility, usability, and engagement among diverse stakeholders, including patients, caregivers, and healthcare professionals. This study evaluates the RS Sardjito and RS Jogja websites using six key dimensions: Attractiveness, Perspicuity, Efficiency, Dependability, Stimulation, and Novelty. Both websites excel in Attractiveness and Perspicuity, showcasing visually appealing and user-friendly platforms. RS Sardjito demonstrates strength in Efficiency, enabling effective task completion, while RS Jogja outperforms in Dependability and Novelty, reflecting higher reliability and innovation. However, areas for improvement include Novelty and Dependability for RS Sardjito and Efficiency for RS Jogja, with both platforms requiring enhancements in Stimulation to deepen user engagement through interactive features. These findings offer actionable insights for driving policy development in healthcare website design and functionality, addressing key areas such as accessibility, usability, efficiency, reliability, and innovation. Policies should prioritize user-centered design principles, implement robust security measures to strengthen reliability, and encourage creative approaches to foster innovation. Additionally, regular benchmarking and user feedback mechanisms should be institutionalized to ensure continuous improvement. By systematically addressing these dimensions, healthcare organizations can optimize digital platforms to improve access to healthcare services, enhance patient engagement, and advance the overall quality of healthcare delivery, contributing to the growing body of research on healthcare website optimization and aligning user experience with organizational goals
Developing a Delphi Validated Instrument for Assessing Digital Forensics Readiness Based on COBIT 2019
Full text linkThe increasing complexcity of cyber threats has reinforced the need for robust digital forensic readiness in higher education institutions. However, existing frameworks often lack integration between forensic capabilities and IT governance practices. Objective: This study aims to develop and validate a new instrument to assess digital forensic readiness based on the COBIT 2019 framework. Methods: A three-round Delphi process was conducted with seven digital forensics and IT governance experts to develop and validate a new instrument comprising forty proposed indicators across six domains. Result : The instrument achieved full context, with I-CVI values increasing from 0.60 to 0.99 and IQR values reaching 1.00 across all items. Implications: The validated instrument integrates governance and forensic principles, providing a standardized tool for institutional self-assessment and policy development, while contributing methodologically through the use of a structured Delphi validation process
Sentiment Analysis of Tokopedia Customer Reviews Using BiLSTM and IndoBERT with Comparative Analysis of Preprocessing and Labeling Methods
Full text linkThis study addresses key challenges in Indonesian sentiment analysis related to preprocessing, labeling strategies, and class imbalance. It compares the performance of BiLSTM and IndoBERT using user reviews collected from Tokopedia. The dataset was manually and automatically labeled, then processed under three preprocessing schemes. Both models were trained with tuned hyperparameters and imbalance-handling techniques and evaluated through twenty rounds of stratified five-fold cross-validation. Performance was assessed using balanced accuracy and F1-score. IndoBERT achieved the highest results, with balanced accuracy up to 0.85 and F1-scores up to 0.83, while BiLSTM reached balanced accuracy up to 0.78 and F1-scores up to 0.76. Applying class weight and focal loss improved model performance by approximately 2% to 11% over the baseline. BiLSTM demonstrated greater training efficiency, requiring only 1 to 2.5 minutes per epoch, compared with IndoBERT’s 2.6 to 3.6 minutes. Although manual labeling remained superior in capturing contextual nuance and emotional cues, GPT-based labeling showed strong agreement with the human annotations. A four-way ANOVA revealed that all main factors and several interactions significantly influenced classification outcomes. Overall, BiLSTM provides faster training efficiency, whereas IndoBERT delivers higher predictive accuracy