Volume 14 | Issue 3 |

 Abstract

The unchecked spread of misinformation across online platforms continues to threaten informed public discourse, electoral integrity, and social cohesion. This study investigates the effectiveness of five distinct machine learning and deep learning approaches for automatically identifying false news content. The classifiers evaluated include Logistic Regression, Multinomial Naive Bayes, Support Vector Machine (SVM), Random Forest, and a fine-tuned BERT transformer model, all benchmarked on the publicly available LIAR dataset. A standardized text preprocessing pipeline comprising tokenization, stop-word filtering, and TF-IDF vectorization was applied to traditional models, while raw token sequences were fed directly into the BERT tokenizer. Empirical results demonstrate that the BERT-based model achieves the highest classification accuracy of 96.4%, substantially surpassing conventional baselines. The findings establish BERT as the most effective approach while also affirming the practical utility of lighter-weight models in resource-limited deployment scenarios.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Fake News Detection, Misinformation, Text Classification, Natural Language Processing, BERT, TF-IDF, Machine Learning.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The increasing accumulation of fruit and vegetable waste has become a major environmental concern worldwide. One effective approach to manage this waste is to utilize this waste for the synthesis of bio-enzymes. Bio-enzymes are complex organic substances composed of protein chains (enzymes), organic acids, and mineral salts produced through fermentation of fruit and vegetable waste or their peels. Conventional chemical cleaners and sanitizers often contain harmful chemicals such as chlorine, ammonia, and synthetic surfactants that may cause environmental pollution and health issues including skin irritation and respiratory problems. In contrast, bio- enzyme based organic sanitizers are biodegradable, non-toxic, and environmentally friendly. Therefore, organic bio-enzyme cleaners/sanitizers are considered a safer and sustainable alternative to harmful chemical cleaners for household sanitation and cleaning purposes.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Bio-Enzymes; Multipurpose; Citrus Peel Waste; Fermentation; Organic waste management; Eco-friendly sanitizer.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This work proposes a deep learning-based traffic monitoring system using YOLO to detect vehicles and classify traffic as low, normal, or heavy. It also identifies possible accidents and displays results on a real-time dashboard, improving traffic management and road safety.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Traffic Monitoring, YOLOv8, Deep Learning, Vehicle Detection, Traffic Density Estimation, Flask, Smart City, Real-Time Monitoring

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Muga Silk, known as the golden thread of Assam is the rare, ecofriendly, exquisite and expensive, durable fabric which has been emerged as the source of income and sustainable livelihood for rural communities, specially for women. This product tagged with geographical indication (GI) promotes export and cultural heritage of Assam. Muga Silk farming is not free from challenges. To strengthen this industry, government of Assam has taken various initiatives through financial support, setting up infrastructure, introducing modern techniques in production etc. This paper aims to study the production trend of muga silk in Assam along with income and employment generation. It is also tries to highlight some issues and remedial measures taken by the government.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Muga Silk, production, employment, income

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

"Yog ka itihaas"

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

"Yog ka itihaas"

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

MockMeUp - Interview Coach is a system designed to help students and job seekers practice interviews and improve their skills. It provides a mock interview environment where users can answer questions and receive live feedback on their performance. The system evaluates responses and suggests improvements, helping users build confidence and prepare effectively for real job interviews.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Mock Interview , Interview Coach , Artificial Intelligence (AI), Natural Language Processing (NLP), Answer Evaluation , Feedback System , Machine Learning , Interview Preparation , User Performance Analysis , Web Application

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Currently, Buddhism is one of the foremost religions in the World, which gained a significant following within India since its revival in the mid-20th century by foremost scholars like Dr. B.R. Ambedkar, Bhikkhu Jagdish Kasyap and Anagarika Dhammapala etc. Within the academic field, specially in Indian social and humanities studies curriculum, Buddhist philosophy, history as well as the art and architectures achieved a special place. With excavation of Buddhist sculptures, icons and stupas, along with remains of mah?vih?ras from regions of Gandhara, Mathura, Orissa, Ajanta etc. regions, studies on its architectural style and artistic significance, has been gaining interest of several scholars and art enthusiasts. In the present article, the inception and development of Buddhist sculpture making, along with the various ceremonies for crafting such idols has been discussed.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This project focuses on building a GPT-based intelligent code review system to support programming education by providing clear, accurate, and meaningful feedback to students. Instead of checking only whether the program output is correct, the system understands the logic, structure, and intent of the code, similar to how an experienced mentor would review it. It analyzes aspects such as correctness, readability, efficiency, and common logical mistakes, and explains errors in a way that helps students learn rather than just fix the code. The feedback is designed to guide learners step by step, without directly revealing final solutions, encouraging independent thinking and problem-solving. By adapting its explanations based on the student's skill level, the system offers a more personalized learning experience. This approach reduces dependency on manual evaluation and enables consistent, scalable, and high-quality code reviews. Overall, the system aims to bridge the gap between automated assessment tools and human teaching, making programming learning more effective and learner-friendly.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

GPT-based code review, intelligent code review system, programming education, automated feedback, code analysis, machine learning, natural language processing, AI in education, personalized learning, code quality assessment, error detection, logical error analysis, readability and efficiency, adaptive learning system.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Industrial safety remains a critical concern in manufacturing and construction environments worldwide. Despite stringent regulations and advanced protective equipment availability, workplace accidents continue to occur due to non-compliance with Personal Protective Equipment (PPE) usage. This paper presents "InduSafe Vision" - an intelligent, automated worker-aware industrial safety and compliance monitoring system that leverages cutting-edge computer vision and deep learning technologies. The proposed system utilizes YOLOv8 (You Only Look Once version 8) for real-time PPE detection, integrated with QR code-based worker identification, automated voice alerts, and comprehensive reporting mechanisms. The system continuously monitors video feeds from workplace environments, detects workers, identifies them through QR codes on their identity cards or helmets, checks for PPE compliance, and immediately alerts violators through personalized voice warnings. All violation incidents are documented with photographic evidence and stored in a SQLite database for future reference and analysis. The system architecture comprises seven core modules: camera capture, PPE detection using YOLOv8n lightweight model, QR code scanning via OpenCV, voice alert generation through pyttsx3, evidence capture, database management, and a web-based dashboard built with Flask. The implementation demonstrates significant improvement over manual monitoring systems by providing continuous, unbiased surveillance with instant feedback. Key features include real-time violation detection at 30 FPS processing speed, QR code-based automatic worker identification with <100ms recognition time, immediate voice alerts with <2 seconds latency, and automated email reporting for both supervisors and workers. The system achieves high accuracy in detection (>85% with pre-trained YOLOv8n model) while maintaining computational efficiency suitable for deployment on standard consumer hardware. Experimental results demonstrate that the system successfully detects PPE violations with high reliability, reduces the workload on safety supervisors by approximately 50%, and creates a comprehensive digital record of safety compliance. The automated reporting system ensures accountability and provides data-driven insights for safety improvement initiatives. The modular architecture allows for easy integration with existing security camera infrastructure and scalability across multiple work zones. The system represents a practical implementation of Industry 4.0 principles, demonstrating how advanced technologies can address real-world industrial challenges while remaining accessible and cost-effective. InduSafe Vision embodies cyber-physical systems integration, Internet of Things (IoT) connectivity, real-time data processing, and decentralized decision-making.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Industrial Safety, PPE Detection, YOLOv8, Computer Vision, Deep Learning, QR Code, Worker Identification, Voice Alert, Automated Monitoring, Flask, SQLite, Industry 4.0, Real-time Surveillance.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The rapid expansion of digital trading platforms has made online trading more accessible; however, users continue to face challenges related to secure transactions, real-time data handling, and efficient management of trading activities. Issues such as maintaining data consistency, ensuring secure user authentication, and providing seamless real-time updates remain critical concerns in online trading and marketplace systems. To address these challenges, this paper presents the technical design and implementation of TradeNova, an online trading and marketplace platform developed as a major project. The system is designed to provide a secure, scalable, and user- friendly digital environment for buying, selling, and managing trade transactions in real time. The platform architecture is built using modern web technologies to support efficient data handling, authentication, and transaction management. This paper details the system architecture, implementation methodology, and functional workflow, demonstrating how TradeNova ensures data integrity, responsive performance, and reliable user interaction across different modules.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Artificial Intelligence, Machine Learning, Stock Market, Trading

  License

Creative Commons Attribution 4.0 and The Open Definition