Authors - Siqabukile Ndlovu, Ernest Mkandla Abstract - This paper presents a novel approach to improve test case prioritisation in continuous integration environments by integrating semantic features derived from test case descriptions. The approach uses a Convolutional Neural Network (CNN) model that considers both structured and semantic features. The target variable for prioritisation is defined using a threshold on calculated priority values, helping to identify the most critical test cases. We further engineer interaction and polynomial terms with the semantic feature to capture complex relationships. These polynomial terms allows the model to capture non-linear relationships. For example, a small change in the semantic feature might not matter much in the middle range, but a large implicit_prob (e.g., above 0.9) might have a disproportionately strong effect on the priority. Through a 5-fold crossvalidation, our results demonstrate that while the semantic features do not statistically significantly improve overall F1-score, they yield a statistically significant enhancement in Average Percentage of Fault Detection (APFD) (pvalue = 0.0055). This indicates that incorporating semantic understanding enables the model to effectively reorder test cases, leading to earlier detection of critical faults and increased efficiency in testing efforts within live CI/CD environments.
Authors - Juhi Patel, Tejaskumar Bhatt Abstract - Growing realization of the need for sustainable agriculture has seen Artificial Intelligence (AI) applications implemented in precision agriculture, advocating for sustainable use of resources and enhanced yields in crops. The authors present an AI-driven model that employs multi-modal data of soil attributes, weather, and vegetation indices to predict crop yield. It develops a model with satisfactory test prediction accuracy based on starting from a Linear Regression model (MSE of 0.0835 and R² score of 0.9714). This shows that the model can potentially capture linear trends for big meteorological change influence on crop yields. Feature correlation analysis determines NDVI and soil moisture as the most significant predictors, with the important roles of monitoring vegetation health and effective water management yielding better agricultural results. Although Linear Regression is a robust starting point, this research also paves the way for integrating sophisticated AI techniques like Neural Networks and Ensemble Learning models to tackle non-linear relationships and interactions. In comparative analysis, we have confirmed the utility of interpretable models for scalability, reliability and applicability in smallholder farming. Implications: The result fills the void between the abstract AI structures and the usable application, providing a widely replicable artifact to improve decision-making in precision agriculture. In addition to prediction accuracy, the study also emphasizes on scalability and ethical consideration, the latter coupled with the area of application of IOT based AI empowered systems and data privacy. Further studies will seek to evaluate the model in different agricultural regions, explore new machine learning methods, and implement real time decision support systems. This study contributes to global efforts towards achieving sustainable agriculture, ensuring food security, and limiting the environmental impact of agriculture by providing knowledge that can be used directly.
Authors - Bhumi Patel, Mann Patel, Aum Mehta, Nishat Shaikh, Priteshkumar Prajapati Abstract - Deepfake generation techniques have advanced rapidly in recent years, posing significant risks for misinformation and privacy. In this paper, we propose an ensemble-based deepfake detection framework that leverages EfficientNet-B4 as the backbone architecture for both image and video analysis. By incorporating attention mechanisms and siamese training strategies, our system enhances feature discrimination and improves robustness against subtle manipulation artifacts. The framework is trained and evaluated on two widely adopted benchmarks—the Deep- Fake Detection Challenge (DFDC) and FaceForensics++ datasets. Experimental results demonstrate that the ensemble approach outperforms individual models, achieving higher accuracy and improved log-loss metrics, while also providing interpretability via attention maps. We further discuss the integration of temporal consistency analysis to better handle video data, and outline future directions for real-time deepfake forensic systems.
Authors - Putri Haryani, Sinatria Arka Daniswara, Randany Zevanya Sihombing, Gavikal Hanif Pasopati, Ridho Rambu Bassae, Andhika Hendra Adi Wibisono Abstract - This study investigates the implementation of the Institutional Financial Application System (SAKTI) in digitalizing public sector accounting within Indonesia’s National Research and Innovation Agency (BRIN). Recently, BRIN is a newly formed government institution resulting from the merger of several institutions. It manages budgets from various sources and operates within a complex organizational structure. BRIN faces significant challenges in achieving accurate and timely financial reporting. Using a qualitative case study method, data were collected through semi-structured interviews and observations at BRIN’s Bureau of Planning and Finance. The findings reveal that SAKTI, alongside the MONSAKTI monitoring system, plays a pivotal role in supporting accrual-based financial reporting in compliance with Government Regulation No. 71/2010. Despite progress in automation and integration, several challenges persist, including limited system performance, partial manual processing, and constrained operational budgets. The study highlights the strategic importance of aligning digital infrastructure, internal controls, and human resource capacity to sustain accountability and reporting quality. These insights are expected to inform similar public institutions undergoing digital financial transformation amid structural complexity.
Authors - Dina Sekar Vusparatih, Handy Martinus, Eshaby Mustafa, Ahmad Hidayat Ahmad Ridzuan Abstract - Malaysia has become the primary destination for Indonesian patients over the past two decades. However, international medical tourism carries significant risks due to limited information, restricted communication access caused by distance and language barriers, while the industry heavily relies on trust for critical health-related decision-making. This study focuses on how Island Hospital implements technology strategies to reduce uncertainty in making informed decisions. A qualitative approach using a case study method was employed to gather the necessary data. Interviews were conducted with Indonesian patients and their family members at the hospital, as well as with Indonesian government representatives in Penang, Malaysia. To enrich the findings, participant and documentation observations were also applied. The results indicate that the hospital integrates Human-Computer Interaction (HCI) using four approaches in designing its web-site/portal and mobile application. This enables patients to communicate with healthcare providers despite existing barriers, and supports them in planning both pre- and post-treatment procedures.
Authors - Gweneaella Lyrika R. Aguinaldo, Nellisa F. Cortez, Jocelle Marie S. Dador, Kathleen Kaye L. Getonzo, Loran Ann G. Gonzaga, Angela B. Navarro, Remelyn J. Vicente, Manuel J. Logatoc, John Kenneth M. Arcayos Abstract - This study examines how key store attributes influence shoppers’ purchase intentions and, ultimately, customer satisfaction in five selected supermarkets in Cavite, Philippines. Store attributes—defined as accessibility and cleanliness, product assortment, promotion, price, and customer relations—are hypothesized to drive the likelihood that a customer will buy (purchase intention) and feel satisfied. Employing a quantitative correlational design, the researchers ad-ministered an adaptive, 45-item Likert-scale questionnaire (4-point) both face-to-face and online to 380 supermarket patrons. Instrument reliability was con-firmed via Cronbach’s alpha coefficients between 0.8740 and 0.9379 (good–excellent). Respondent demographics indicated a predominantly young (18–28 years, 75.5 %) and female (62.4 %) sample, with over half (51.6 %) reporting a monthly income below ₱10,000. Descriptive analysis (frequency, percentage, weighted mean, and standard deviation) revealed that accessibility and cleanliness received the highest effectiveness rating (mean = 3.44, SD = 0.43), while customer relations scored lowest. Inferential analysis using Spearman’s rank-order correlation showed a moderate positive relationship between purchase intention and customer satisfaction (ρ = 0.549, p < 0.001). Among individual attributes, promotion correlated most strongly with purchase intention (ρ = 0.507), and price exhibited the highest link to customer satisfaction (ρ = 0.567). These findings suggest that supermarkets seeking to boost sales, and loyalty should prioritize promotional strategies to drive purchase intent and calibrate pricing structures to enhance satisfaction. Optimizing accessibility, cleanliness, and other in-store factors can further improve customer experience, yielding competitive advantage in the retail sector.
Authors - Pradnya H Desai, Parikshit Mahalle, Pankaj Chandre Abstract - The exponential growth of the Internet of Everything (IoE) demands secure, intelligent access control mechanisms to manage data flow and device interaction efficiently. This paper proposes a novel architecture that integrates post-quantum cryptography, federated learning, and explainable AI to ensure privacy- preserving, real-time decision-making for resource-constrained IoE environments. The edge nodes, equipped with lightweight cryptographic engines and context-aware training modules, perform preliminary data processing and secure communication via GG-ULL and PQCrypto protocols. The cloud leverages federated learning to train distributed models and uses a multi-agent policy engine for dynamic access decisions. Blockchain-backed audit trails ensure accountability, while the explainable AI module enhances transparency in access control. The system supports mission-critical services like SCADA and health data lakes, guaranteeing secure and interpretable decision flows. This architecture paves the way for resilient, scalable, and intelligent access control frameworks suitable for next-generation IoE ecosystems.
Authors - Deepika K M, Rohith H P, Srinivas D B, Lakshmi H Abstract - The need for real-time object recognition is growing in a few applications, including robotics, surveillance, and autonomous vehicles. Modern object recognition technique YOLOv5 achieves high accuracy while maintaining real-time performance. This paper proposes a real-time, highly accurate object recognition method using YOLOv5. The system, which was created using PyTorch and Python, is trained and evaluated using the COCO dataset. The proposed system enables fast object detection and achieves outstanding precision and recall rates using single-shot detector architecture. Additionally, the detection accuracy is greatly improving with the introduction of YOLO and its architectural descendants. YOLOs are frequently employed in a variety of contexts, mostly because of their speedy conclusion rather than due to the accuracy of their detection. The YOLO detection accuracy, for instance, ranges between 63.4 and 70. The suggested system is ideal for real-time object detection applications since experimental findings demonstrate that it performs better than current object detection systems in terms of accuracy and speed.
Authors - Asra Fatema Zakir Baig, Amol Mashankar Abstract - Hospital leadership styles are crucial determinants of employee performance and patient satisfaction. In this research, secondary research findings are integrated to explore the influence of transformational, transactional, and laissez-faire leadership styles on healthcare delivery. Transformational leadership is always linked with greater staff engagement, job satisfaction, and better patient care outcomes. Transactional leadership is less consistent in its findings, typically enhancing short-term efficiency but with little developmental impact over the long term. Laissez-faire leadership has typically been associated with negative organizational performance because of the lack of direction and responsibility. The literature highlights the requirement for healthcare managers to embrace transformational styles as a means to create a safe work environment as well as an improved patient experience. Emotional intelligence, communication skills, and encouraging teamwork should feature prominently in any future leadership programs aimed at enabling sustainable healthcare greatness. (Alilyyani, B., Wong, C. A., & Cummings, G. G., 2018), Effective hospital leadership has a profound impact on the performance of staff as well as patient satisfaction. The current paper integrates evidence from research studies to explore how the transformational, transactional, and laissez-faire styles of leadership influence healthcare. The review emphasizes the significance of leadership in creating an optimistic workplace culture, improving staff participation, and finally enhancing patient care.(Wong, C. A., & Cummings, G. G. (2013))
Authors - Najah Najmia Halim, Okta Bayu Prihatma Putra Abstract - This paper uncovers the system-level challenges confronting freelancers in emerging markets, framing freelancing as a form of digital entrepreneurship with significant social implications. Through a systematic literature review of 23 peer‑reviewed studies (2020–2024) using the PRISMA protocol, we identified five critical challenge domains: opaque algorithmic management that limits freelancer autonomy, financial instability due to irregular income and weak social protection; regulatory ambiguity that erodes trust in digital platforms, technical barriers including limited infrastructure and digital literacy, and career development gaps marked by burnout and isolation. These challenges disproportionately affect freelancers in developing regions, highlighting structural inequalities within the global digital labor market. By synthesizing literature across geography and sector, this study positions freelancing not just as a labor trend, but as a pressing issue of digital inclusion and social sustainability. The findings inform policymakers, platform designers, and support institutions about intervention points to foster equitable, resilient freelance ecosystems. Future research must explore how technology, policy, and cross-sector partnerships can create inclusive innovation frameworks that sustain independent digital work.
