http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/36 2026-06-07T23:53:57Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/17232 Zero-shot Learning for Predicting Unseen Student Activities in Educational Platforms Shanto, Emon Islam Zero-shot learning (ZSL) has emerged as a groundbreaking approach in machine learning, enabling models to classify unseen categories by leveraging the relationships between known and unknown categories through semantic knowledge. Within educational platforms, accurately predicting and understanding student activities is vital for personalizing learning, tracking progress, and improving adaptive learning systems. However, the wide diversity of potential student activities makes collecting labeled data for every scenario unfeasible, posing a significant limitation to traditional supervised learning methods. This study addresses this limitation by introducing a ZSL framework specifically designed to predict unseen student activities. The proposed framework leverages semantic embeddings, such as word2vec and BERT, to establish meaningful connections between known and unknown activities, enabling accurate predictions without labeled examples. The framework is thoroughly evaluated using real-world datasets of student interaction logs, with performance assessed across metrics such as accuracy, precision, recall, and F1-score. The results highlight the framework's ability to deliver strong predictive performance while providing valuable insights into the relationships between activity categories. By bridging the gap between labeled and unlabeled data, this research showcases the transformative potential of ZSL in advancing educational platforms. It demonstrates how ZSL can enhance adaptive learning systems, foster student engagement, and equip educators withactionable insights, driving the development of smarter, more personalized educational technologies. Project Report 2025-01-19T00:00:00Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/17231 E-Commerce Web Development Project for Tech Shop IT Uzzaman, Moner An e-commerce website serves as a digital platform for buying and selling products or services over the internet. It enables businesses to reach a global audience, offering features such as product catalogs, secure payment systems, and order management. Customers can browse, compare, and purchase items conveniently from anywhere, while businesses benefit from streamlined operations and enhanced market presence. By integrating user-friendly interfaces, personalized experiences, and efficient delivery systems, e-commerce websites drive innovation and convenience in modern retail. Project Report 2025-01-11T00:00:00Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/17230 Coffee Shop Management System Saimum, Iftekhar Mohammad This project focuses on developing a coffee shop website that simplifies customer interactions and enhances the overall experience. The platform enables users to view menus, order items, book tables, read blogs, and subscribe to newsletters, making their engagement convenient and efficient. Administrators benefit from features that allow them to manage menus, process orders, oversee table reservations, and communicate with customers via email. By integrating intuitive design with practical functionality, the website improves operational processes and elevates customer satisfaction, addressing the growing demand for digital solutions in the hospitality sector. Project Report 2025-01-19T00:00:00Z http://dspace.daffodilvarsity.edu.bd:8080/handle/123456789/17229 A Hybrid Machine Learning Model for Enhanced Prediction of Gestational Diabetes Using Diverse Datasets Miti, Ramisa Maliyat Gestational diabetes mellitus (GDM) is a significant health concern affecting maternal and fetal well-being, necessitating early and accurate predictive models. This study presents a novel hybrid machine learning model integrating Random Forest, Support Vector Machine, and Gradient Boosting Machine through a stacking ensemble approach. The hybrid model achieved superior performance across two datasets, with accuracy scores of 92.7% and 89.02%, significantly outperforming individual models. The integration of diverse data sources, including clinical, biochemical, and demographic variables, enhanced the model's robustness and generalizability. Metrics such as precision (91.5% and 86.05%), F1-Score (92.3% and 73.18%), and ROC-AUC (0.94 and 0.91) underscore the model's ability to balance precision and recall effectively. The study addresses key research gaps, including generalizability issues, data integration, and scalability. By incorporating hyperparameter tuning, model pruning, and quantization, the hybrid model is optimized for deployment in resource-constrained settings, demonstrating scalability and efficiency. Despite its promise, challenges such as the need for external validation across diverse populations and addressing biases in training data remain. Future research should focus on fairness-aware algorithms and longitudinal studies to ensure equitable healthcare outcomes. This hybrid model showcases its potential as a reliable tool for early GDM detection, enabling timely interventions and improving maternal and fetal health outcomes. Its integration into clinical workflows and adaptability across healthcare settings highlight its significance as a step forward in precision medicine. Thesis Report 2025-01-11T00:00:00Z