East Delta University

Purpose This study aims to investigate the role of Artificial Intelligence (AI) in enhancing environmental, social and governance disclosures in organizations in Bangladesh, focusing on the integration of sustainability culture (ESC) and climate resilience (CLR) within organizations. Design/methodology/approach Using a cross-sectional survey of 434 respondents from various industries in Bangladesh, this study uses structural equation modeling to test the hypotheses on the relationship between AI integration, ESG disclosures and organizational change. Findings The findings show that AI integration significantly enhances ESG reporting, fostering a strong sustainability culture and improving organizational climate resilience. The study also reveals that stakeholder awareness mediates the relationship between ESG disclosures and sustainability culture, with Organizational innovation performance moderating the effect of AI use on sustainability culture. Practical implications The study provides valuable insights for organizations aiming to enhance their ESG reporting practices through AI, offering practical strategies for improving sustainability culture and climate resilience in response to global climate challenges. Social implications The study highlights the role of digital transparency and stakeholder consciousness in advancing sustainability goals. It emphasizes the importance of ethical AI use and inclusiveness in environmental reporting, helping firms balance economic growth with societal well-being and climate justice in vulnerable regions. Originality/value This pioneering study in an emerging economy empirically validates a moderated mediation model integrating AI with ESG accounting, advancing stakeholder theory and the resource-based view in organizational change contexts.

Abstract We explore the spillover among ASEAN-5 stock markets (Malaysia, Singapore, Thailand, the Philippines, and Indonesia) and the categorical economic policy uncertainty (EPU) of three major economies (China, Japan, and the United States), employing an R 2 decomposed connectivity approach. Our empirical findings depict significant connectedness among ASEAN-5 stock markets and the categorical EPUs of three major economies, where instantaneous temporal spillovers are greater than lagged spillovers. We find that most categorical EPUs are shock receivers, whereas the ASEAN-5 stock markets are shock transmitters. Chinese monetary and exchange rate policy uncertainties, along with Japanese fiscal policy uncertainty, act as net shock transmitters. However, the roles of categorical EPUs and ASEAN-5 stock markets vary across total, contemporaneous, and lagged transmissions. We also find that the GFC, the Euro debt crisis, the U.S.‒China trade war, COVID-19, and the Russia–Ukraine conflict caused dynamic fluctuations in shock transmission. We also show that oil price changes, global financial stress, climate change, and gold market ups and downs are key factors that link economic policy uncertainties and the ASEAN-5 markets in different time frames. Hence, our empirical results provide valuable insights for managing risks and formulating strategies for addressing economic policy uncertainty and world crisis events.

Cardiovascular diseases remain one of the leading causes of mortality worldwide, particularly in low- and middle-income countries. Early and accurate prediction of heart disease is essential for timely intervention and improved patient outcomes. Machine learning techniques offer promising solutions; however, challenges such as class imbalance, lack of interpretability, and limited real-world validation persist. In this study, a machine learning–based heart disease prediction framework was developed using a real-world clinical dataset comprising 5000 patient records collected from healthcare facilities in Bangladesh. Data preprocessing included cleaning, feature encoding, train–test splitting, and class imbalance handling using the Synthetic Minority Oversampling Technique (SMOTE). Multiple machine learning models—Logistic Regression, Decision Tree, Support Vector Machine, and Random Forest—were evaluated using 10-fold stratified cross-validation. Model performance was assessed using accuracy, precision, recall, and F1-score. SHAP (SHapley Additive exPlanations) was employed to enhance model interpretability. The best-performing model was deployed as a web-based decision support system. Among the evaluated models, the Random Forest classifier achieved the best performance, with an accuracy of 98%, recall of 96%, and F1-score of 96%. Ablation studies demonstrated the effectiveness of SMOTE, feature integration, and ensemble modeling. SHAP analysis identified clinically relevant features contributing to heart disease prediction, enhancing transparency and trust in model decisions. The proposed framework provides an accurate, interpretable, and practical solution for heart disease prediction using real-world clinical data. The integration of explainable machine learning and web-based deployment highlights its potential for clinical decision support. Future work will focus on multi-center prospective validation and adaptive model updating to further improve generalizability and real-world applicability. • A large real-world clinical dataset of 5000 patients was used for robust heart disease prediction. • A Random Forest model with SMOTE achieved high performance (98% accuracy, 96% recall). • Comprehensive ablation and model comparison analyses validate methodological robustness. • SHAP-based explainability provides transparent and clinically interpretable predictions. • The proposed system is deployed as a web-based clinical decision support tool.

Abstract Software as a Service (SaaS) startups have an extremely high failure rate. Even though many business owners are adapting at creating goods, they frequently fail because they create goods that consumers do not genuinely want. It is essential to validate an idea before investing in its development. However, existing approaches, such as manual research or basic keyword tools are frequently out of date and unable to keep up with the rapid changes in social media public trends. We present the Adaptive SaaS Idea Validator, a novel system that employs artificial intelligence to forecast the success of a startup idea, in order to close this gap. Using the PRAW library, we gathered actual Reddit discussions to create a custom dataset. The sentiment (positive or negative emotions) and engagement levels of these conversations were then determine using a program called DistilBERT. To determine which AI model could most accurately predict success , we compared several different models. Conventional models that achieved accuracy included Random Forest (93.66%), Gradient Boosting (97.12%), and 1 LightGBM (97.12%). On the other hand, our suggested Offline Reinforcement Learning approach yielded the best results (97.23%). This method, in contrast to static models, makes better judgments about which concepts are feasible by learning from past data. We anticipate that this study will assist new business owners in lowering risk. Before devoting time and resources to creating the finished product, they can use this tool to test their concepts against actual social data.

Student-centered pedagogy has gained immense popularity among the educators and learners of the educational system for its’ theoretical compatibility and diverse learning options. Educational researchers in Bangladesh urge that educational institutions should adopt SCL to enhance the competence of their students to cope with the 21st century job market. In Bangladesh this pedagogy is unfamiliar and vague to the majority. Teacher dominated classroom culture is so intrinsic that the instructors and students can hardly perceive the norms of SCL. This research was undertaken to explore how much teachers perceive the conception of SCL. There were five research questions to measure the level of perception and qualitative approach was adapted to elicit the subjective view of the teachers. Data were collected through semi-structured interview using six open-ended questions and data were analyzed using thematic analysis method. The findings revealed that teachers of the college are well-perceived regarding some aspects of SCL such as the role and responsibilities of teacher and learner, benefits, potentiality and challenges of implementing SCL. The educators also felt that Student-Centered Learning (SCL) can significantly address the issue of student absenteeism. At the same time, the study suggested that teachers are not sufficiently acquainted with the aspect of power relationship between teacher and students and inconsistent assessment procedures of TCL and SCL. It is expected that the findings of the study will instigate all to take necessary measures to bring potential changes. Finally, as the research was a small-scale study and represented teachers’ subjective view, it is recommended that future research in this regard should adopt a mixed method approach for better outcome.

The purpose of the study is to examine the hotel employees’ perceptions of adopting renewable energy in Bangladesh’s hospitality industry, emphasizing the mediating role of management support. The necessity of this study stems from the hospitality sector’s significant energy consumption and its potential to contribute to sustainable development. Despite global momentum, renewable energy adoption in Bangladesh remains limited, especially in hospitality. This study addresses this gap by exploring key psychological and organizational factors influencing adoption. Data was collected from 360 hotel employees using a structured questionnaire and analyzed via SmartPLS. Findings reveal that self-effectiveness perception, environmental concern, and renewable energy cost significantly influence the intention to adopt renewable energy, while management support plays a crucial mediating role. The study offers practical recommendations for policymakers and industry stakeholders to enhance sustainable energy practices, contributing to both theoretical understanding and practical implementation of renewable energy in emerging economies.

This comprehensive review identifies gaps in current frameworks, such as limited support for real-time compliance auditing and challenges in scaling security measures for large-scale applications. The evolution of full-stack development frameworks has revolutionized the software development landscape, enabling seamless integration of front-end and back-end functionalities. With increasing reliance on web and mobile applications, security and compliance have become critical focal points in framework design and implementation. This review examines advances in full-stack development frameworks, emphasizing their security and compliance models to address modern-day challenges such as data breaches, unauthorized access, and regulatory non-compliance. The study explores prominent full-stack frameworks, including MEAN (MongoDB, Express.js, Angular, Node.js), MERN (MongoDB, Express.js, React, Node.js), and Django, highlighting their inherent security features. These include robust authentication mechanisms, encryption protocols, and defense against common threats such as SQL injection, cross-site scripting (XSS), and cross-site request forgery (CSRF). Furthermore, the review delves into compliance models embedded within these frameworks, focusing on their adaptability to global standards like GDPR, HIPAA, and PCI DSS, which govern data privacy and security. Emerging trends such as the integration of artificial intelligence (AI) for anomaly detection, serverless architecture for reducing attack surfaces, and blockchain for enhancing data integrity are also discussed. The adoption of DevSecOps practices in full-stack frameworks has further streamlined the embedding of security measures during the development lifecycle, mitigating risks proactively. Recommendations for future research include enhancing framework modularity to incorporate evolving security technologies, leveraging AI for predictive threat modeling, and ensuring compatibility with emerging compliance mandates. By presenting a thorough analysis of existing and emerging security and compliance paradigms, this review serves as a valuable resource for developers, researchers, and organizations aiming to enhance the reliability and trustworthiness of full-stack applications in an increasingly digital and regulated environment.

Purpose This study aims to examine the impact of institutional pressures on green ship recycling practices (GSRP) and investigates the moderating role of resource availability on the relationship between institutional pressures and GSRP. Design/methodology/approach This study collected a total of 257 top- and mid-level management personnel associated with the ship recycling industries as a respondent under the purposive sampling method. The authors used PLS-SEM and artificial neural network (ANN) to evaluate the hypotheses’ relationships. Findings The results found that all the institutional pressures (i.e. mimetic, normative and regulatory pressure) have significant positive impacts on GSRP. Additionally, this study’s results indicate that resource availability moderates the relationship between mimetic pressures and GSRP. Moreover, the findings reveal that normative pressures are the most important factors in predicting GSR behavior, followed by regulatory pressures, mimetic pressures and resource availability. Originality/value This study is innovative as it examines the relationships between institutional pressures and GSRP in the ship recycling industry, using PLS-SEM and ANN techniques. It also advances the knowledge of resource availability and GSRP, contributing to resource-based view theory.

Roads are the nation's leading transportation system, but the global road network is only 39% paved. Moreover, potholes on pavements pose a significant safety hazard, leading to high maintenance costs and road accidents. Manual detection of potholes proves inefficient and costly, requiring extensive human labor before repair activities begin. Therefore, there is a need to create automated detection systems for potholes. Merging computer vision with deep learning approaches, offering a better alternative for real-time pothole detection adeptly. The study utilizes 9,240 annotated pothole images, divided into 66% for training, 23% for validation, and 11% for testing, to train and evaluate YOLO11n, YOLO11s, YOLO11m, YOLO11l, and YOLO11x models. YOLO11s delivered the best mAP50 score of 0.848, with a precision of 0.887 and recall of 0.778, surpassing other models, making YOLO11s the best match for real-time pothole detection. This research is one of the first to evaluate YOLO11’s detection capability, which assures the possibility of improving pothole detection and maintenance strategies, reducing costs and accident probabilities.

Lung cancer (LC) is one of the leading causes of cancer-related morbidity and mortality worldwide. Inflammation is a driver of cancer initiation and progression, affecting processes such as angiogenesis, antiapoptotic pathways, and DNA adduct formation. Cytokines are small proteins that can accelerate or slow tumor growth by controlling associated signaling processes such as cell proliferation, metastasis, and apoptosis. This review reveals the role of tumor necrosis factor-alpha (TNF-α), interferon-gamma (IFN-γ), transforming growth factor-beta (TGF-β), and interleukins in LC. Macrophages play a role in non-small cell lung cancer (NSCLC) pathogenesis and are associated with poor prognosis. A nested case–control study revealed that elevated concentrations of IL-6 and IL-8 were strongly associated with the risk of LC. Specifically, the odds ratio (OR) for IL-6 and IL-8 in former smokers (fourth quartile vs. first quartile) was 2.70 (95% confidence interval [CI], 1.55–4.70) and 2.83 (95% CI, 1.18–6.75), respectively. Because C-reactive protein levels are elevated in patients with NSCLC with larger and higher-grade tumors, CRP has been identified as a systemic indicator of chronic inflammation. Insulin-like growth factors influence cellular signal transduction pathways and contribute to tumorigenesis. Soluble tumor necrosis factor receptors have been explored for their role in NSCLC prognosis, highlighting their association with chromogranin. Transient receptor potential cation channel, subfamily M, member 7 (TRPM7), urokinase plasminogen activator, matrix metalloproteinases, and monocyte chemoattractant protein-1 have been identified with a focus on their expression patterns and prognostic significance in LC tissues. Moreover, lung angiogenesis induces vascular endothelial growth factor, soluble intercellular adhesion molecule-1, myeloperoxidase, and tissue inhibitors of metalloproteinase expressions. In conclusion, this review thoroughly summarized the inflammatory cytokines and specific factors influencing LC, providing the basis for further research on potential treatment approaches.