Authors - Truong Cong Hau, Nguyen Quoc Cuong, Tran Tat Dat, Duong Ngoc Anh, Nguyen Son Duong Abstract - In an era of increasing environmental awareness and social accountability, ESG (Environment, Social, and Governance) practices have become a strategic priority for logistics enterprises in Southeast Asia. This study aims to model the multivariate impacts of ESG implementation on firm sustainability, with a particular focus on the mediating role of Time-Critical Logistics (TCL). ESG integration is expected to enhance environmental performance, social responsibility, and governance transparency, while TCL supports the efficiency of time-critical logistics operations. Using a quantitative approach and Structural Equation Modeling (SEM), the study analyzes data from 369 from major logistics enterprises in Southeast Asia. Findings are anticipated to confirm the positive influence of ESG on sustainability, with TCL playing a key mediating role.
Authors - Tam Nguyen Minh, Tai Vu Thanh, Quynh Chi Truong, Thi Ai Thao Nguyen Abstract - The growing reliance of enterprises on internal and information systems has led to an increased risk of cyber attacks, which pose serious threats to the confidentiality, integrity and availability of data. Intrusion Detection Systems (IDS) are essential for identifying malicious activities; however, conventional IDS approaches often struggle with evolving and sophisticated attack vectors. Recent advances in machine learning (ML) offer new opportunities to enhance IDS capabilities through adaptive and data-driven models. This study proposes a hybrid architecture that integrates convolutional neural networks (CNN), long-short-term memory (LSTM), and gradient boosting machines (GBM), leveraging both supervised learning and advanced feature representation techniques, addressing key challenges such as limited labeled data, diverse traffic patterns, and resource constraints in small and medium enterprises (SMEs). The proposed approach includes an end-to-end pipeline for data collection, preprocessing, feature extraction, and data labeling, combined with aug-mentation techniques to improve model generalization. Experimental evaluation across different network scales demonstrates the effectiveness of the method in detecting both known and novel attack types, while maintaining computational efficiency. The findings contribute to a practical and scalable solution for enter-prise network security, with implications for real-world IDS deployment and fu-ture research in ML-driven cybersecurity.
Authors - Niranjan C Kundur, Anil B C, Sreenatha M, Jayasimha S R Abstract - A hybrid machine learning framework designed to predict burnout levels and employee retention among IT professionals in urban India a sector increasingly challenged by mental health issues, high attrition, and work-life imbalance. The proposed framework integrates Gradient Boosting Regression for continuous burnout score prediction and Random Forest Classification for binary retention outcomes. Data were collected from 300 IT employees across five metropolitan firms, encompassing cross-sectional and temporal variables related to work conditions, wellness policies, and demographic profiles. Rigorous preprocessing, feature encoding, and 5-fold cross-validation ensured robust model performance. The burnout prediction model achieved an R² score of 0.87 and a Mean Absolute Error (MAE) of 4.12, while the retention classifier yielded an accuracy of 85%, F1-score of 0.82, and AUC-ROC of 0.88. Key predictors included flexible hours, remote work options, and job satisfaction, with gender emerging as a significant moderating factor—female employees demonstrated 12% greater sensitivity to burnout-related outcomes. The findings highlight the utility of predictive analytics in identifying high-risk employees and guiding personalized wellness and retention strategies. This hybrid framework offers HR leaders a scalable, data-driven tool to optimize workforce well-being and reduce attrition in the evolving landscape of Indian IT workplaces.
Authors - Albrecht Kati, Vilko Jyri Abstract - Restaurants and their customers value natural food products (NFPs), and demand for them is growing steadily due to product qualities such as taste and freshness, as well as support for sustainable development. However, the availability of natural food products from suppliers can be challenging. Availa-bility challenges mean that the market for natural food products is inefficient. This also makes it difficult for restaurants to define their competitive strategies. Local restaurants are often small, entrepreneur-driven businesses with a few dozen employees. In small restaurants, resources are spent on daily operations, and systematic long-term development is often given less attention. However, systematic operations require clear, planned customer segmentation and, through this, the development of competitive strategies as a means of creating value. The main objective of the study is to increase understanding of the link between cus-tomer segmentation and competitive strategies in business and the added value they generate. In addition, the study aims to guide restaurants towards more sys-tematic business development. The data was collected through qualitative re-search and in-depth interviews with restaurant owners and managers in Finland. The research was based on thematic analysis, and the results were documented using NVivo software. Restaurants were found to have insufficient knowledge of how NFTs can be used to create value-added business and how customers can be segmented using competitive strategies.
Authors - Ervine de Vasconcellos, Tapiwa Gundu Abstract - The rising acceptance of digital technology to improve patient care has made cybersecurity one of the biggest issues for healthcare workers. Having an urgent need to fully understand accepted cybersecurity standards and the main causes of non-compliance. Knowing how healthcare workers see cybersecurity policies becomes essential as a result of the following interruptions in services and more complex assaults that target hospitals and clinical environments. The objective of the research was to provide guidelines for addressing the non-com-pliance with cybersecurity policies by healthcare workers in South Africa. A lit-erature review was carried out to determine the current status of cybersecurity policies in healthcare and a systematic literature review was done to determine the causes of healthcare workers non-compliance with cybersecurity policies. The systematic literature review provided some themes which were then analysed and guidelines were created based on them.
Authors - Jonalyn Echano, Carla Dayne N. Ferreras, Jhaniella Anne T. Ganuelas, Maria Angela P. Obsuna, Regienald G. Obina Abstract - As digital commerce continues to thrive, the Philippine government has implemented the Creditable Withholding Tax on Electronic Marketplace Transactions to ensure tax compliance among online sellers. This study explores how online sellers in Quezon City cope with the challenges brought by this regulation. Specifically, it examines their coping mechanisms in relation to the complexity of tax laws, perceived fairness of the tax system, quality of tax admin-istration, effectiveness of enforcement measures, and the level of penalties and sanctions. A quantitative research design was employed, using a structured survey distributed via email and social media. The study gathered responses from 113 online sellers operating within Quezon City. Statistical tools such as frequency distribution, weighted mean, and ANOVA were used to analyze the data. Findings revealed that while most sellers perceived the tax system as generally fair and the administration as responsive, they struggled with the complexity of the laws. Many sellers acknowledged the effectiveness of enforcement and the influence of penalties in motivating compliance. However, differences in coping mechanisms were significantly evident only when sellers were grouped by busi-ness experience and perceived fairness—while factors like capitalization and business category showed no major influence. The results highlight the need for clearer guidelines, simplified tax processes, and continuous education efforts. This research contributes to a deeper understanding of how small digital entre-preneurs adapt to evolving tax policies, offering insights for policy improvements that support voluntary compliance and fairness in the growing e-commerce sec-tor.
Authors - Anand Ghugare, Atharva Rajoba, Dinesh Kumar Saini Abstract - Quantum computers poses a significant threat to traditional cryptography systems, which depend on integer factorization. Theoretically, Shor’s algorithm, a quantum polynomial-time algorithm, allows quick factorization of huge semiprimes, compromising the RSA protocol. This paper shows hardware-based execution by implementing Shor’s algorithm using IBM’s quantum resource and Qiskit framework to factor the number 221. The study examines the success probability and performance limitations of implementation on noisy quantum hardware. Our results provide valuable insights into effective quantum cryptanalysis and pave the path for scaling such approaches for larger numbers.
Authors - Alma Delia Torres-Rivera, Laura Alma Diaz-Torres, Mario Leonardo Nieto Antolinez Abstract - This study proposes a theoretical framework for integrating artificial intelligence into sustainable business models for urban public transport. Drawing on a literature review and two in-depth expert interviews (in data analytics and operations), it identifies key structural challenges in the sector, including fragmented concession models, rising costs, declining demand, environmental pressures, and weak alignment with national policies such as the National Policy for Urban Collective Public Transport. The findings indicate that artificial intelligence alone is insufficient to achieve sustainability; it must be integrated with robust enterprise architecture and adaptive strategic planning. To support this integration, the study presents a six-phase roadmap: structural diagnosis, enterprise architecture redesign, strategic planning, AI integration, training and change management, and continuous evaluation. This roadmap offers a practical and ethical pathway for digital transformation aligned with the Sustainable Development Goals. Ultimately, artificial intelligence can improve efficiency, resilience, and sustainability—provided its deployment is context-sensitive and supported by institutional leadership, organizational capacity, and cross-sector collaboration.
Authors - Markus Heikkila, Alisa Misukka, Ilkka Kero, Kati Albrecht, Jyri Vilko Abstract - Entrepreneurship in the natural product field is growing, founded in the utiliza-tion of sustainable natural resources and entrepreneur’s personal values and motives. Natural product field offers interesting perspectives for entrepreneurship where business activity is shaped by financial targets, personal value system and relationships with nature. The field’s immatureness, heterogeneity, and undeveloped supply chains highlight the decisions independent entrepreneurs make regarding their business strategy. This research explores the path towards entrepreneurship of two entrepreneurs within the natural product field regarding entre-preneurship motivation and commercialization strategies in the form of an extreme case study. It addresses that entrepreneurship within natural product field is driven by passion, customer entered approach, and willingness to combine per-sonal values with business activity. Case companies differ from one another re-garding their willingness to grow, educational background, and brand building. However, similarities can be found regarding valuing quality, customer feedback, and self-directed information seeking. The results expand previous understanding about entrepreneurship motives, indicating how strategic make or buy decisions are tied into entrepreneur’s identity, values and resource acquisition. Specifically in natural product entrepreneurship, commercialization is not only a technical process but rather based on entrepreneur’s value system, covering product origin, authenticity and ethical grounds.
Authors - Nasir Ansari, Amit Kumar Sharma Abstract - A significant challenge in human-computer interaction is enabling effective communication for individuals with hearing impairments. This research addresses the need for accessible solutions by implementing a neural network-based system for recognizing American Sign Language (ASL) letters from images captured via a standard webcam. The primary objective is to reduce the communication barrier between the deaf and hearing communities by leveraging machine learning techniques for hand gesture classification. A Convolutional Neural Network (CNN) was utilized for both feature extraction and classification tasks. The model was trained on a publicly available Kaggle dataset comprising 87,000 images across 29 ASL alphabet classes. Experimental results demonstrate that the proposed model achieves high classification accuracy, indicating its potential applicability in real-time ASL translation systems. When we sit down to plan the next phase of the project, well add real-time gesture tracking, mix in voice and touch data, and shrink the whole model so it runs smoothly on phones and tablets.
Authors - Saptarshi Paul, Pratul Kalita, Abu Ahmed Waliullah Mazumder, Saurav Paul Abstract - With the rise of multilingual technology and the growing demand for inclusive language tools, there remains a significant gap when it comes to resources and research dedicated to Assamese, one of the major languages spoken in North east India. The core objective is to bridge that gap by leveraging recent advances in deep learning—specifically the Transformer architecture—to build a translation model that is both accurate and adaptable. The system is implemented using the OpenNMT-py framework which supports flexible training of encoder- decoder models with attention mechanisms. To train the model, a parallel corpus consisting of approximately 124,000English-Assamese sentence pairs was sourced from the Samanantar dataset available on Kaggle. The data was prepared using a machine translation (MT) preprocessing script, which utilizes Sentence Piece to perform sub word tokenization. This step ensures that rare and compound words are broken down into more frequently occurring subunits, thereby improving the model’s ability to learn patterns in low-resource language settings. The trained model was evaluated using the BLEU score, which is used to compare the overlap between machine-generated outputs and human reference translations. The evaluation results indicated that the Transformer-based approach provides promising performance, particularly in producing fluent and contextually relevant translations, despite the limited size of the dataset.
Authors - Homayoun Rabiei Pour, Jyri Vilko Abstract - Non-timber forest products (NTFPs) are wild products harvested from the forest, except timber, there’s a significant potential for application in the food supply chains. This systematic literature review evaluates the process potential of various NTFPs based on the feasibility and technological aspects of these un-derutilized natural resources for food processing. By using Web of Science, pub-lications were retrieved with a specific search strategy of keywords mainly con-sisting of food, non-timber forest products, processing, and cost. After screening based on defined inclusion criteria, relevant articles were selected for analysis. The results contained different applications of NTFPs such as wild fruits, shea butter, bamboo shoots, and plant-derived polysaccharides in food manufacturing, preservation, and nutritional enhancement. Different processing methods such as drying, extraction, fermentation, and enzyme treatment were documented with corresponding evaluations of safety, functional properties, and economic viabil-ity. While interest in sustainable and local foods is increasing, there are also large gaps related to scale-up feasibility, market integration, and cost analysis. The re-view concludes with recommendations for future research, particularly empha-sizing innovation in food-grade processing and value-added product develop-ment from NTFPs.
Authors - R. Kowsalya, S. Ravi, C. Manusha, A. Saranya, T. Kalaichelvi Abstract - A significant obstacle to early skin cancer detection is the diversity of skin types and colors. Compared to convolutional neural networks (CNNs), Vision Transformers (ViTs) are a powerful alternative. By applying self-attention mechanisms to model global dependencies within an image, ViTs can capture long-range contextual information more effectively than traditional CNNs that use local convolutions to capture spatial patterns. A study describes how Vision Transformers perform better on large-scale image datasets than CNNs when trained with enough training data to achieve maximum performance. Additionally, hybrid architectures are investigated, which combine the advantages of CNNs and transformers strategies to improve generalization and training efficiency. It will help researchers learn about the latest deep-learning techniques and comparative analysis. It will also be helpful to know which type of Transformer is most suitable for accurately detecting the early stage of skin cancer classification and detection.
Authors - Shane Maluleke, Tebatso Gorgina Moape, Ernest Mnkandla Abstract - The exponential growth of consumer and service provider reviews on digital platforms has generated substantial big data online, creating both opportunities and challenges for business analytics. Businesses often use these reviews to evaluate customer satisfaction, service quality, and overall brand perception. Due to the vast amount of data generated, traditional analysis methods are often inadequate for efficient processing. Hence, most companies employ sentiment analysis techniques to analyze substantial volumes of data from reviews. Sentiment analysis is a subset of natural language processing that enables the automatic classification of text-based feedback according to the emotional tone expressed. In this paper, sentiment analysis is conducted on the Uber driver app, a ride e-hailing service, within the South African context. This study intentionally focused on driver reviews instead of customer reviews, as most research predominantly focuses on passenger satisfaction, service quality, and pricing strategies, while driver perspectives remain understudied. The methodology employed in the paper involved a systematic data mining process, followed by text pre-processing, thematic code analysis of the collected data, and the application of Naïve Bayes and Random Forest algorithms to classify the Uber driver app reviews. The Random Forest model outperformed the Naïve Bayes classifier, with an accuracy of 0.9023, while Naïve Bayes achieved an accuracy of 0.8333.
Authors - Mathew Philip, Ruksana Banu, Amala Nirmal Doss Savari, Mimy Mathew, Ahmed Aboualam Abstract - The rapid progress of Internet of Things (IoT) technology and digitalisation is transforming the logistics sector into more efficient operational processes and data-informed decision-making. This shift is driven by evolving market demands, sustainability imperatives, and the emergence of innovative business models. This study aims to identify and analyse current research trends in the application of IoT in smart logistics management, and to understand how these technologies address existing research gaps in traditional logistics systems. The study presents a bibliometric analysis and examines the concept of smart logistics. It also reviews and analyses IoT-based smart Logistics Management (SLM) research, demonstrating how IoT influences smart logistics. The study reveals that the most prominent area, based on Scopus research from 2017 to 2025, was IoT, which is regarded as a key revolution. Terms like efficiency, visibility, and responsiveness highlight the operational priorities IoT technologies seek to optimise within logistics and supply chain contexts.
Authors - Pornpimol Chaiwuttisak Abstract - This study developed predictive models for the closing prices of five leading technology stocks: GOOGL, MSFT, AAPL, NVDA, and META by employing five advanced machine learning and deep learning techniques: Light Gradient Boosting Machine (LightGBM), Extreme Gradient Boosting (XGBoost), Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU), and Long Short-Term Memory (LSTM). The modeling framework integrated sentiment scores derived from financial news articles specific to each stock us-ing the VADER Sentiment Analysis tool, in conjunction with a range of macro-economic indicators. Model performance was evaluated separately for each stock using Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) as primary metrics. To determine whether statistically significant dif-ferences existed among the predictive performance of the models across all stocks, the Friedman test was employed, followed by the Wilcoxon signed-rank test for post-hoc pairwise comparisons. The empirical results indicated that XGBoost achieved superior predictive accuracy for MSFT and AAPL, GRU outperformed other models for NVDA and META, while RNN yielded the most accurate forecasts for GOOGL.
