International Journal of Advances in Data and Information System
Not a member yet
161 research outputs found
Sort by
Identification of Drug Material Melting Conditions from Hot-Stage Microscopy Images Using Active Contour and Support Vector Machine Methods
Full text linkThe hospital pharmacy installation plays an essential role in ensuring the quality of pharmaceutical supplies. One important stage in drug production is raw material analysis, particularly melting point determination as a purity indicator. Conventional methods, such as capillary tubes, are limited in accuracy and prone to subjectivity. This study aims to develop an automated image-based monitoring system integrated with Hot Stage Microscopy (HSM) to objectively detect real-time morphological changes in pharmaceutical materials. The system was designed using digital image processing stages consisting of image acquisition, processing, and output. Images were captured using a binocular microscope and processed on an Odroid XU4 mini-computer. Phase boundaries were identified using the Active Contour segmentation method, while texture features were extracted using the Gray Level Co-occurrence Matrix (GLCM) at four orientation angles. Classification was performed using a Support Vector Machine (SVM) with a Radial Basis Function (RBF) kernel. The results showed that the Active Contour method effectively detected melting phases, and the SVM achieved an accuracy of 91.67%, precision of 91.89%, sensitivity of 91.67%, and an F1-score of 91.66%. The system successfully distinguished pure Paracetamol from mixtures with Gallic Acid and Ferulic Acid
Systematic Literature Review - Turning Heads: Quantifying Hedonic, Eudaimonic, and Behavioural Engagement in 360° Tours
Full text linkThis systematic literature review (SLR) examines the relationship between 360° virtual experiences, user engagement, and conversion outcomes. It explores how immersion influences emotional, cognitive, and behavioral responses across fields such as education, tourism, cultural heritage, retail, and gaming. Drawing from empirical studies published between 2013 and 2025, the review investigates how factors like presence, perceived control, enjoyment, and cognitive load mediate engagement in immersive environments. Experimental and quasi-experimental designs dominate the analyzed studies, often using frameworks such as PLS-SEM to assess mediation and moderation effects. Findings show that interactivity (hotspots, mini-maps, gamification) and guidance mechanisms (narration, AI support) tend to enhance behavioral engagement, while high visual fidelity, contextual relevance, and user-centered control strengthen both hedonic and utilitarian conversions, including purchase intentions and revisit likelihood. However, gaps remain in methodological rigor, including small samples and inconsistent engagement metrics. The review concludes that effective 360° virtual experiences integrate emotional engagement with usability to transform immersion into measurable outcomes. It recommends future research to emphasize cross-domain comparisons, standardized measures, and longitudinal studies to better understand how immersive systems sustain engagement and influence conversion behavior over time
Integration of Machine Learning and GAP Analysis for a Data Driven Lecturer Performance Evaluation System
Full text linkThe objective of this research is to design and implement a performance evaluation system that combines Machine Learning for data processing, predictive modeling, and pattern recognition with the GAP method to measure discrepancies between expected competencies and actual performance. Eight primary criteria were cooperation, communication, initiative, alertness, discipline, leadership, problem solving, time usage each consisting of several sub-criteria. The study involved 18 lecturers, and the evaluation was conducted using a web-based decision support system equipped with machine learning models trained to classify performance levels and identify underlying patterns within the assessment data. System usability was examined through four categories: ease of use, completeness, accuracy, and interface composition. The results show that the integrated system successfully identified the highest-performing lecturer (Lecturer 7) with a score of 6.1801, followed by Lecturer 12 with 4.9314 and Lecturer 4 with 4.1157. Usability testing also yielded positive outcomes, with scores of 89% for ease of use, 87% for completeness, 90% for accuracy enhanced through machine learning validation and 88% for interface composition. These results produced an overall average of 88%, classifying the system as Very Worthy. In conclusion, integrating Machine Learning and GAP Analysis in a web-based DSS significantly improves the effectiveness and efficiency of lecturer performance evaluation. The system accelerates data processing, enhances assessment quality, and strengthens decision-making through predictive analytics and automated classification. This framework offers a valuable reference for future performance evaluations in higher education institutions seeking accountability, transparency, and data-driven decision-making
Uneven Transitions in Container Ship Capacity Across Indo-Pacific Economies (2010–2022): Integrating PCA, ANOVA, and Clustering Evidence
Full text linkWe examine uneven transitions in container ship capacity (TEU per ship) across five Indo-Pacific economies China, Singapore, Australia, Vietnam, and Indonesia during 20102022 using an integrated statistical framework that combines ANOVA, Welch ANOVA, GamesHowell post-hoc tests, Principal Component Analysis (PCA), and clustering. Results reveal persistent divergence: China and Singapore maintain high-capacity fleets (>10,000 TEU/ship), Australia stabilizes in the mid-tier range (~7,000 TEU/ship), while Indonesia and Vietnam experience rapid but low-level growth (<6,000 TEU/ship). ANOVA confirms significant cross-country differences (F=28.33; p<0.001; 0.65), with Welch ANOVA yielding consistent results under unequal variances (p<0.01). PCA indicates one dominant component (PC199.5%) explaining most variance, forming three readiness clusters: high, medium, and low capacity economies. These patterns suggest that policy inertia, infrastructure bottlenecks, and green transition constraints drive the uneven capacity development. The study contributes by introducing TEU per ship as a cross-national indicator for maritime readiness, linking statistical divergence to SDG targets 8, 9, 10, 13, and 14, and offering empirical guidance for low-carbon fleet transition and port modernization in emerging economies.
Data Analytics Capability Maturity and Governance Gaps in PT ABC: A Diagnostic Case Study
Full text linkThis study provides a within-case diagnosis of data analytics capability maturity in PT ABC to inform refinement of its 2025 to 2026 governance roadmap. Using a qualitative-first mixed-method diagnostic case design (QUALQUAN), this study conducted five semi-structured key-informant interviews first, followed by a cross-sectional survey of 76 purposively selected analytics-involved employees, analyzed using the Shortened TDWI Data Analytics Maturity Model (STDAMM). The overall maturity mean was 3.31 on a 5-point scale (bootstrap 95% confidence interval: 3.15 to 3.46), classifying the assessed respondent pool as Level 3 (Established). Dimension means were tightly clustered (3.27 to 3.35) with substantial confidence-interval overlap, suggesting that respondents perceive maturity as a coupled system rather than sharply separated capability domains. A small descriptive pattern places analytics use and infrastructure marginally above data management and organizational enablement, consistent with interview accounts of definition inconsistency, limited socialization of dictionary and stewardship practices, and recurring manual reconciliation of figures across systems. The findings support a governance-first refinement of the roadmap that prioritizes standardization and ownership mechanisms to improve the comparability and decision-readiness of analytics outputs, while treating subgroup comparisons cautiously where measurement comparability is not supported
A Problem-Driven User Experience Model for Evaluating Government Transparency Platforms: Evidence from A Regional Command Center
Full text linkThe rapid expansion of digital transparency initiatives has encouraged local governments to adopt data-driven platforms to enhance public accountability. However, many transparency platforms struggle to achieve sustained public engagement due to unresolved user experience issues. This study proposes and empirically validates a problem-driven user experience evaluation model for government transparency platforms by focusing on three UX problem dimensions: Extra Time or Effort, Unexpected Experience, and Evolving Limitations. Using survey data from 147 valid users of the Command Center website of Sumedang Regency collected between October and December 2025, the model was tested using Partial Least Squares Structural Equation Modeling. The results indicate that all three UX problem dimensions significantly influence user experience satisfaction and continued intention to use, both directly and indirectly, with satisfaction acting as a partial mediator. The findings demonstrate that reducing cognitive effort, ensuring experiential consistency, and addressing systemic limitations are critical for sustaining public engagement with transparency platforms. This study contributes to the e-government and UX literature by offering a problem-oriented evaluation framework that emphasizes structural usability frictions rather than interface aesthetics, providing actionable insights for improving digital governance and public transparency
The Role of Simplified Enterprise Architecture (Mini TOGAF) in Improving Project Management Governance and Decision-Making
Full text linkThe development of information technology in the digital age requires companies to have a system architecture that is aligned with their business strategy. One commonly used framework is The Open Group Architecture Framework (TOGAF). However, the complexity of TOGAF is often an obstacle to its implementation, especially for organizations with limited resources. This study introduces a Mini TOGAF framework—an adaptive simplification of TOGAF 10 artifacts—designed specifically for digital creative enterprises. Unlike previous simplification models that mainly addressed SMEs in traditional industries, this framework integrates agile principles and stakeholder-centered validation cycles, reflecting the current evolution of enterprise architecture practice in 2023–2025. The method used is Design Science Research (DSR) with three main cycles: the Relevance Cycle to identify organizational needs, the Rigor Cycle to review relevant theories and methods, and the Design Cycle to iteratively design and evaluate artifacts. Data was collected through interviews, observations, and literature studies, then validated by the company. The results of the study show that the application of Mini TOGAF can improve architectural understanding, operational efficiency, business agility, and corporate strategy alignment. The simplification of TOGAF artifacts has been proven to reduce the complexity of implementation without reducing the main benefits of the framework. This study contributes to enterprise architecture literature by proposing an adaptive TOGAF 10 simplification model that strengthens the theoretical link between architectural governance and digital business agility. These findings provide practical contributions for organizations in adopting Enterprise Architecture efficiently and adaptively to modern business needs
Usability Evaluation and Alternative Interface Design Recommendations for the GoFood Merchant Application
Full text linkThe GoFood Merchant application serves as a critical platform for merchant onboarding and product activation within Indonesia’s digital food delivery ecosystem. Despite its key role, user feedback indicates persistent usability issues, particularly in corporate onboarding. This study evaluated and improved the usability of GoFood Merchant by focusing on workflow complexity, information clarity, and interface consistency. A mixed-methods exploratory case study was conducted involving quantitative and qualitative data collected from merchant owners, merchant staff, and IT experts. Quantitative data were obtained using the system usability scale (SUS) questionnaire, while qualitative data were collected through task-based usability testing, open-ended questions, and an evaluation based on heuristic design principles derived from Shneiderman’s eight golden rules of interface design. The initial evaluation of the existing (“as-is”) interface yielded an SUS score of 63.83, indicating usability below the accepted benchmark, and revealed 10 consensus usability issues related to complex task flow, unclear system feedback, and inconsistent interface elements. Based on these findings, a redesign was developed using user-centered design (UCD) principles through iterative low-fidelity and high-fidelity prototyping. The redesigned (“to-be”) prototype was subsequently evaluated using the same instruments. Usability was significantly improved, with an SUS score of 85 and average principle-based evaluation ratings above 4.0 across all interface design principles derived from Shneiderman’s eight golden rules. These findings demonstrate that systematic usability evaluation combined with iterative UCD-based redesign can effectively enhance user satisfaction, efficiency, and comprehension in complex onboarding systems. This study provides validated design recommendations for improving GoFood Merchant onboarding and contributes empirical evidence for usability-driven interface redesign in large-scale mobile platforms
Spatio-Temporal AIS Big Data Analytics of Vessel Traffic Patterns in Kaohsiung Port
No full textMaritime traffic management in major ports requires a comprehensive understanding of vessel movement patterns to ensure operational efficiency and safety. This study presents a spatio-temporal analysis of vessel traffic in Kaohsiung Port, Taiwan, utilizing a 10-month snapshot of AIS data (December 2024–October 2025). Employing quantitative methods including Kernel Density Estimation (KDE) for spatial intensity mapping, grid-based discretization for traffic density quantification, and temporal resolution analysis at multiple scales, the research identifies key operational hotspots and peak traffic periods. The analysis encompasses 1,247,890 AIS records from diverse vessel types, revealing distinct spatial clustering patterns in port entrance channels, anchorage zones, and terminal areas. Temporal analysis demonstrates pronounced diurnal and weekly cyclical patterns, with peak traffic intensities occurring during daytime operational hours and weekdays, reflecting commercial shipping schedules and port operational rhythms. The KDE-based hotspot identification reveals high-density zones concentrated within 0.5 nautical miles of major container terminals, indicating critical areas requiring enhanced traffic monitoring and collision avoidance measures. Grid-based traffic density quantification provides granular insights into vessel distribution across different port sectors, enabling zone-specific risk assessment and resource allocation strategies. The findings reveal complex spatio-temporal patterns that reflect the port\u27s role as a major container hub in the Asia-Pacific region. Despite data quality limitations such as unspecified vessel types (59.9%) and incomplete destination fields, the results provide actionable insights for port authorities to enhance safety, optimize operations, and support strategic planning. This methodological framework demonstrates scalability and transferability to other port environments, contributing to the advancement of data-driven maritime traffic management system
Performance Evaluation of AdamW, RMSProp, and Nadam Optimizers on EfficientNetB2 Model for Image Data Classification
Full text linkThis study examines the effect of different optimization algorithms on the performance of the EfficientNetB2 model in classifying lung and colon histopathology images. Three commonly used optimizers AdamW, RMSprop, and Nadam were analyzed to compare their influence on convergence trends, classification accuracy, and overall learning consistency. Using a five-class dataset covering benign and malignant tissue samples, the experimental results show that all three optimizers are able to deliver reliable predictions, although with varying performance characteristics. RMSprop emerges as the most effective optimizer, achieving the highest accuracy across all evaluation stages, with 99.05% during training, 99.16% on validation, and 98.72% on testing, along with the lowest loss values. This indicates that RMSprop facilitates faster and more stable convergence compared to the other two methods. AdamW also demonstrates strong predictive performance but shows limitations when distinguishing cancer types with closely similar morphological structures. Nadam attains high accuracy in early stages yet exhibits lower initial stability than RMSprop. Overall, pairing EfficientNetB2 with RMSprop provides the most optimal configuration for this classification task. These results offer valuable insights for designing better training strategies and strengthening the effectiveness of medical imaging based computer aided diagnostic systems