Authors - Sayeda Rahnuma Akthar, Fatima Ibrahim, Muhammad Rezaul Islam, Marzan Binte Hasan, Md. Ataur Rahman, Zinia Hafiz, Farzana Sadia, Mahady Hasan, M. Rokonuzzaman Abstract - This study explores the challenges faced by Bangladeshi software Small and Medium-sized Enterprises (SMEs) in adopting ISO/IEC 26550 for Software Process Improvement (SPI). Data collected from 12 SMEs revealed uneven implementation across key process areas, with high familiarity in domain requirements (up to 95\%) but limited engagement in testing and validation practices. To address these gaps, the study proposes a lightweight PDCA (Plan–Do–Check–Act) framework tailored to the unique resource constraints of SMEs. Un-like traditional models such as CMMI or ISO/IEC 12207, the proposed approach emphasizes simplicity, cost-effectiveness, and role-based adaptability, making it more practical for SMEs in emerging economies. The framework supports scalable and gradual process improvement, enhances quality assurance, and facilitates stronger alignment with ISO/IEC 26550 in resource-limited environments.
Authors - Sushmitha, Vijayalaxmi Abstract - This paper provides prediction of air quality using computational intelligence techniques. The study employs various Computational Intelligence techniques to evaluate the data associated with environmental pollution and hence to predict the air quality focusing on PM2.5 and PM10 concentrations. The performance of each model is assessed with standard evaluation metrics. Through a series of simulations, it is observed that the results suggest that computational intelligence techniques play an essential role in improving the accuracy of air quality prediction models.
Authors - Jaride Chama, Hajraoui Kenza, Zaidane Kaoutar, Benabdelkrim El Filali Yasmine Abstract - With technological developments and the digitalization of business processes in the banking sector, physical and virtual environments are complementing each other. Electronic channels are no longer considered as additional or less costly way of interacting with customers. Today, they represent a considerable and constantly growing part of the activities of bank institutions world-wide. The aim of this work is to assess the impact of trust in the development of fintech banking through a transnational analysis. The results of this work can be used to map the countries that studied this topic. The findings of this study reveal that, despite the fact that research in this field has been carried out in a wide variety of economic and cultural contexts (45 countries were surveyed), the previous studies agreed on the strong presence of trust as an entry barrier for the development of financial technology in the banking sector. In addition to trust, other adoption factors appear to be essential for consumers. The vulnerability resulting from digitalization in the banking sector has enabled us to formulate some suggestions for further investigations.
Authors - Md. Abbas Ali Khan, Md. Musfique Anwar, Mohammad Hanif Ali, Touhid Bhuiyan, Md. Ataur Rahman Abstract - Personalized driving assistance, security, and usage-based insurance models all depend on driver identification in intelligent transportation systems. Conventional techniques for identifying drivers primarily rely on biometric or manually engineered feature data, which may not adequately capture the nuances of unique driving actions. Using time-series driving data, including vehicle speed, steering angle, and acceleration, we present a unique contrastive learning-based method for driver identification in this study. The suggested model uses contrastive learning to compare driving sessions from the same driver to those of other drivers to learn how to distinguish between them. Rather than requiring manual feature extraction, our method automatically extracts strong, high-dimensional representations of driving behavior. Concerning driving sequences, the contrastive learning framework aims to minimize the similarity between sequences from different drivers and maximize the similarity between sequences from the same driver. When applied to situations with various driving behaviors and a small amount of labeled data, our strategy outperforms more conventional classification techniques in accuracy. The suggested model offers a scalable and effective solution for real-world applications, as demonstrated by experimental findings on driving publicly available datasets. This model also greatly increases driver identification performance. We applied seven machine learning classification models. The Random Forest Classifier obtained the highest accuracy of 93.57%.
Authors - Ronald C. Barriga, Dan Jeward Rubis, Arlene A. San Pablo, Helinor Y. Medina Abstract - Content video marketing has become vital in the world of marketing for the past few years, in which companies are into it using content creators to expand brand awareness. The use of different social media platforms makes all this possible technological advancement like 5G connectivity that makes the connection faster, the continuous and fast innovation in the electronic industry has made everything easy to access. These make content marketing more vital in establishing corporate brand awareness. Brand recall, recognition and dominance through content video marketing has a significant impact to brand awareness, where brand recall essentially a dominant contributor to Brand Awareness. Furthermore, the model Structural Equation Modeling (SEM) shows that for content video marketing to be effective, the quality of each content must be given importance with consideration also to Knowledge Opinion Leader (KOL) and the Electronic Word of Mouth (eWOM) to strengthen brand awareness. Lastly, SEM positively indicates how significant the three factors mentioned are to brand awareness. The SEM model has a very good fit for the data, explaining a large portion of the variance in brand awareness. The results suggest that to improve brand awareness, strategies that enhance brand recognition, recall, and dominance should be prioritized.
Authors - Dev Parekh, Jay Patel, Keshvi Patel, Dhaval Patel, Priteshkumar Prajapati Abstract - This paper presents CyberShieldAI, a novel framework that leverages BERT (Bidirectional Encoder Representations from Transformers) models to detect cyberbullying in online text communications. Our approach harnesses BERT’s contextual understanding capabilities to identify subtle linguistic patterns, emotional undertones, and semantic structures commonly associated with cyberbullying content. By fine-tuning BERT on carefully curated datasets of online interactions, CyberShieldAI can effectively distinguish between harmless communications and those containing bullying, harassment, or other harmful content. The framework analyzes text bidirectionally, capturing crucial contextual relationships that traditional text classification methods might overlook. Experimental evaluations demonstrate that our BERT-based system achieves significant improvements in cyberbullying detection accuracy, sensitivity, and specificity compared to conventional machine learning approaches. This paper details the system architecture, implementation methodology, dataset preparation challenges, and performance metrics that validate the effectiveness of our strategy. CyberShieldAI represents a significant advancement in automated content moderation systems, which can help create safer online environments, particularly for vulnerable user populations, such as children and adolescents.
Authors - Swati Kale, Jyoti Tipale, Shilpa Sonawane, Kshitij Mulay, Gaurav Patil, Nikita Khumkar Abstract - This paper presents a novel multi-modal speech analysis system that integrates deep learning architectures and “Natural Language Processing (NLP)” techniques to address the limitations of traditional speech evaluation approaches. The proposed system combines a custom “Long Short-Term Memory (LSTM)” based neural network for temporal pattern analysis, TF-IDF vectorization for keyword extraction, real-time spectrogram analysis for acoustic feature extraction, and a dynamic content fetching system from Wikipedia datasets. The system demonstrates enhanced accuracy in content relevance detection, real-time processing with minimal latency, and a scalable architecture suitable for diverse ap-plications, including educational, professional, and research contexts. The study highlights the system's ability to bridge the difference in the theoretical abilities of speech analysis and their practical implementation. The proposed methodology contributes to the advancement of speech analysis by addressing the challenges of insufficient integration between acoustic and semantic analysis, static reference materials, and limited scalability.
Authors - Md. Ataur Rahman, Sabrina Alam, Saadia Binte Alam, Mahady Hasan Abstract - Outcomes-based education (OBE) is reshaping engineering education, with an emphasis on complex engineering problem solving, with clear outcomes directly related to industry needs. This study assessed the impact of including complex engineering problem-solving in the engineering curriculum within OBE on student skills. Through qualitative and quantitative research methods, including course curriculum analysis, student feedback, questionnaires, and student achievement. We collect 3 years of data. Data collected over a three-year research period (2022, 2023, and 2024) at Independent University, Bangladesh (IUB), this study assessed the improvement in skills such as critical thinking, application of advanced tools, sustainability awareness, professional ethics, teamwork, communication, project management, lifelong learning, and entrepreneurship. There are barriers to implementation, such as resource allocation, training of faculty members, and curriculum restructuring, with evidence of positive improvements in these categories. This study strongly recommends greater application of advanced engineering problem-solving techniques in OBE to match educational outcomes to professional engineering requirements better, ultimately producing graduates more adept at solving problems in the real world.
Authors - Rohit Yadav, Rohini Bhosale, Rahul Rathod, Prashant Dhotre, Yashraj Salunkhe Patil Abstract - Phishing continues to be a prevalent threat in cybersecurity, using deceptive URLs to undermine user information and system integrity. Conventional machine learning methods, though effective when handcrafted features are used, tend to perform poorly in generalizing against more obfuscated patterns of attacks. This paper presents an effective phishing URL detection system that takes advantage of the contextual strengths of quantized Large Language Models (LLMs)—namely Mistral-7B and Mixtral-8x7B—combined with interpretable gradient boosting models like CatBoost. The LLMs are employed with prompt-based inference for semantic URL evaluation, whereas the CatBoost model employs lexical and statistical features for structured classification. Tested on a balanced dataset of more than 2.28 mil-lion URLs, the system has high accuracy and low false positive rates. SHAP-based visualizations are employed to offer transparent explanations for model decisions. The findings prove the real-world feasibility of integrating effective LLM reasoning with conventional ML in real-time phishing defense systems, particularly in resource-limited settings.
Authors - Aryan G. Kale, Kajal P. Salampuriya, Nitin Prakash Abstract - The ascent of artificial intelligence technologies, particularly the generative application of GANs, has changed the landscape of media production into a milieu that easily allows - creation of hyper-realistic yet deceptive content. While deepfakes have indeed been misused from entertaining, educative, and access provisions, their abuse causes dreadful threats to mental well-being and social harmony. This study explores the psychological effects of deepfakes, such as anxiety owing to identity theft, impersonation, trust erosion in digital media, cognitive dissonance because of the inability to tell the difference between real and fake content, and social isolation due to cyberbullying and non-consensual pornography.On the social front, deepfakes add to the pool of misinformation spreading, fake news politically entering the narratives, and a myriad of ethical dilemmas, adding to general confusion, anxiety, and distrust among the public at large in institutions. The economic designs revolve around hiring specialists to thwart deepfake manipulation and attending to its legal and psychological aftermath.To ameliorating this, the study points out a projective need for strategies of intervention. These would include developing an AI detecting tool that identifies and flags deepfakes, launching public awareness campaigns to increase digital literacy and critical thinking, instituting solid legal scaffolding with added ethical principles, and other support systems including counseling services supporting the victims of deepfake assaults.The study is primarily inclined toward discussing how integrated management approaches shall serve to address the issues faced concerning deepfakes, while the positives remain intact. Future research is necessary to explore the cultural and demographic nuances of deepfake impacts, understand the influence of the COVID- 19 pandemic on mental health in the context of deepfake exposure, and evaluate the effectiveness of emerging prevention and mitigation strategies.
