Authors - Md. Najmus Sakib Sourov, Md. Ataur Rahman, Urmi Ghosh, Sabera Sultana, Mahady Hasan, Md. Tarek Habib Abstract - English Language Anxiety represents a prevalent issue among learners, significantly affecting their capacity to communicate and acquire knowledge effectively. The study uses machine learning models to offer insights into how anxiety affects students' general well-being and academic performance. English language anxiety among rural high school students in Bangladesh is investigated using machine learning techniques to determine its sources and effects on aca-demic performance. Machine learning techniques like k-Nearest Neighbors (k-NN), Support Vector Machine (SVM), Naïve Bayes, Decision Tree, and Gradient Boosting were used to analyze data from several rural schools, including student scores and qualitative insights on their perceptions of English. The results showed that SVM was the most accurate model, with an accuracy of 99.43%. The study reveals major anxiety patterns and predictions, with a focus on the interaction between academic pressure, English complexity, and student mental health. The findings will guide educators and policymakers in designing focused interventions to relieve English subject anxiety and establish a supportive learning environment.
Authors - Satyasa Khadka, Sandhya Baral, Sudip Tiwari, Sharad Kumar Ghimire Abstract - This paper presents a robust Automatic Number Plate Recognition (ANPR) system tailored for Nepali license plates written in Devanagari script. In this paper, a pipelined model was used that integrates YOLO-based models for license plate and character detection, followed by a CNN classifier trained on 34 Devanagari characters. Two publicly available data sets were used that incorporate diverse lighting, fonts, and structural variations. Data augmentation and additional training on embossed plates enhanced the generalizability of the model. The system achieved a recognition accuracy of up to 93%, demonstrating strong performance under real-world conditions and providing a scalable solution for traffic management in Nepal.
Authors - K. Vidyasagar, G. Lakshminarayana, B. Surya Prasada Rao, K. Vijaychandra, P. Suresh babu Abstract - Nanoscience is the emerging technology to strengthen the devices that are more essential to human life. This chapter discussed structural design considerations of a badminton racket. The stress-strain relationship, modulus of elasticity, bending modulus, and moment of inertia are studied to improve the shaft deflections while smashing the shuttle with high force. Carbon nanotubes are elected for badminton rackets to withstand both rapid forward and side deflections. This paper discussed the physical characteristics of armchairs, Zig-Zag Single-walled carbon nanotubes and double-walled carbon nanotubes (DWCNT). The length of the carbon nanotube is fixed to 200 A. The C-C bond length is considered as 1.4210 A. The stretch factor is considered as ‘1’ for both single-walled and carbon nanotubes (SWCNT). And the ellipse ratio is also considered as ‘1’ for both SWCNT and DWCNT. The screw distortion and bending angle are varied from 00 to 900 for single-walled carbon nanotubes, and the physical elastic properties of the SWCNT and DWCNT are investigated. Finite element analysis is used to analyse SWCNT and DWCNT structures. Nano tube modeler is used to simulate the SWCNT and DWCNT structures.
Authors - Sabar Aritonang Rajagukguk, Eileen Heriyanni, Ahmad Fikron Maulida, Enina Putri Abstract - Multi-SKU demand forecasting poses considerable challenges for re-tail operations due to intricate interdependencies among products, seasonal fluctuations, and external market influences. Conventional forecasting approaches often fail to model these complexities effectively, resulting in poor inventory management and financial inefficiencies. This paper proposes a novel Cross-Modal Attention Transformer (CMAT) architecture for addressing multi-SKU demand prediction in retail contexts. The model employs a dual-attention mechanism to capture temporal patterns within individual SKU time series as well as cross-SKU relationships via an inter-product attention layer. It further integrates multimodal external variables such as promotional campaigns, weather data, and economic indicators using dedicated feature encoders. Evaluation on a large-scale retail dataset comprising 15,847 SKUs across 12 product categories shows a 23.4% reduction in Mean Absolute Percentage Error (MAPE) compared to standard transformer models and a 31.2% improvement over traditional statistical methods. The results indicate potential for improving supply chain efficiency and reducing inventory-related costs through more accurate multi-horizon demand predictions.
Authors - Kshitij Gupte, Kunj Patel, Krisha Desai, Dhrumi Patel, Anjali Jivani Abstract - Advanced Persistent Threats (APTs) pose a serious challenge to cybersecurity due to their stealthy and prolonged nature. Detecting these attacks early is critical, yet traditional methods often fail to catch subtle patterns or require extensive resources. This study investigates the effectiveness of several machine learning algorithms including Logistic Regression, Random Forest, Naive Bayes, and K-Nearest Neighbors (KNN) to identify APT behavior from system and network activity data. A significant focus was placed on reducing the complexity of the detection system without compromising accuracy. Dimensionality reduction using Linear Discriminant Analysis (LDA) was applied to transform high-dimensional data into a more manageable format, and data cleansing techniques were used to handle null and infinite values. The results showed that the KNN algorithm not only achieved the highest detection accuracy but also maintained performance even when trained on just 10% of the dataset. This approach offers a lightweight, high-precision method for APT detection that can minimize processing overhead while improving real-time responsiveness helping organizations enhance their security posture without sacrificing efficiency.
Authors - Justin Q. Cabellon, Ashley MJ A. Andrade, Dave P. Bautista, Kenshin Z. Gutierrez, Dorothy Joy B. Nabo, Angel Anne R. Nuarin, Manuel J. Logatoc, Rexieden N. Barrera Abstract - This study examines the perceptions of small and medium enterprises (SMEs) toward United Nations Sustainable Development Goal 13 (Climate Action). Employing a quantitative survey across multiple industries, it captures numeric data to identify patterns in SMEs’ attitudes and evaluates whether demo-graphic factors— length of operation, nature of business, number of employees, or type of ownership—influence these perceptions. Respondents rated statements on a five-point Likert scale (1 = Strongly Disagree to 5 = Strongly Agree) across five business dimensions: product, service, process, network, and model. Results reveal high overall support for climate-action initiatives, with mean scores indicating positive engagement in each dimension. Statistical analysis shows no significant variation in perceptions among SMEs grouped by any demographic criterion, suggesting a shared recognition of climate action’s importance irrespective of firm characteristics. These findings underscore the uniform commitment within the SME sector to integrate climate considerations into core operations. By illuminating SMEs’ supportive stance and the absence of demographic dis-parities, this research highlights the sector’s potential contributions to global cli-mate objectives. Insights from this study can inform policymakers and practitioners aiming to tailor interventions that bolster SME participation in sustainable development and foster resilient economic growth.
Authors - Parikshit Mahalle, Pradnya H Desai, Pankaj Chandre Abstract - With the widespread adoption of AI-as-a-Service (AIaaS), enterprises increasingly rely on cloud platforms to deploy and access powerful machine learning models. This shift offers scalability and cost-efficiency but also introduces significant security and privacy concerns. Protecting the confidentiality of AI models and user data has become a critical requirement in this outsourced computing paradigm. However, AIaaS platforms face numerous threats, including model inversion, membership inference, extraction attacks, and data poisoning. Traditional security mechanisms are insufficient to address the unique confidentiality risks associated with remote AI model hosting and inference. Furthermore, balancing performance, privacy, and deployment feasibility remains a core challenge. This paper proposes a comprehensive confidentiality-focused architecture for AIaaS, integrating multiple cryptographic techniques and secure hardware layers. The design leverages Homomorphic Encryption, Secure Multiparty Computation, Trusted Execution Environments, and Federated Learning with Secure Aggregation, supported by a centralized Management & Compliance layer. Threat vectors are explicitly addressed across client, cloud, and edge layers, while performance and auditability are maintained. In conclusion, our architecture provides a layered and resilient defense against modern AI threats while maintaining functionality and scalability. It offers a blueprint for deploying privacy-preserving AIaaS infrastructures. Future work will focus on optimizing hybrid models and advancing quantum-resistant protections.
Authors - Mridul Das Joshe, Shyja Rafeek S, Aji Sivanandan Abstract - Fake news has become a critical research area due to the expanding spread of misinformation on digital platforms. As social media becomes a main news source for millions, confirming the authenticity of online material is crucial. This research addresses the problem using a multi-label text classification method, which better identifies the complex nature of misinformation that may fall into multiple categories. In our research, we used the LIAR dataset, a bench-mark collection of 12.8k short political statements labeled with six fine-textured truthfulness classes. A transformer-based architecture, notably a fine-tuned BERT model, is proposed including both textual context and metadata to enhance better classification output. The model achieved an accuracy of 0.7351, presenting superior results than those reported earlier research using the LIAR dataset. Confirming the robustness of the approach across different classes with evaluation metrics and per-label precision and recall. The findings illustrate the potency of transformer-based architectures in identifying different degrees of falsehood. This research adds an extensible and practical solution for real-time fake news detection to support information.
Authors - Agna S. Nath, Arun Kumar V N, Aswathi K.B Abstract - This study investigates the role of WhatsApp Commerce—the use of WhatsApp for business transactions—in empowering women-led Micro, Small, and Medium Enterprises (MSMEs) in Tier-3 cities of India, while advancing Sustainable Development Goals (SDGs) 5 (Gender Equality) and 8 (Economic Growth). Despite India’s digital revolution, gender disparities persist, with women leading only 20% of MSMEs and facing barriers like digital exclusion and sociocultural constraints. Grounded in Kabeer’s Empowerment Theory and Davis’s Technology Acceptance Model, the research employs a mixed-methods approach, combining surveys of 400 women entrepreneurs and 20 case studies across four Tier-3 cities (Bareilly, Jalgaon, Dibrugarh, Erode). Key findings reveal that WhatsApp usage intensity correlates with a 27% revenue increase (β=0.27, *p*
Authors - Kingkar Prosad Ghosh, Ankan Roy, Anupam Singha, Tabassum Tajin Ratry, Nafisha Hossain Abstract - Market Basket Analysis (MBA) identifies co-occurring products for pricing, cross-selling, and placement. This paper reviews Apriori and FP-Growth and introduces the RN-Algorithm, a method that focuses on rules with high lift and confidence by measuring pairwise item co-occurrence. In retail-style transactions, RN provides competitive or superior performance in identifying the top rules while remaining simple to distribute. This study examines RN’s complexity, how its runtime scales with the number of unique items and basket width, and offers a practitioner-oriented comparison of algorithms. We also address RN’s limitations (e.g., pair growth in wide baskets) and propose mitigations, including minimum item support, top-K selection, and distributed counting. The paper concludes with actionable guidance on when to prefer RN over FP-Growth or Apriori in modern e-commerce applications.
