LIVE EVENT DETECTION FOR PEOPLES SAFETY USING NLP AND DEEP LEARNING

Abstract

Large public gatherings such as concerts, sports events, political rallies, religious festivals, and cultural celebrations often generate vast amounts of real-time information through social media platforms and online communication channels. Timely detection of incidents, emergencies, and unusual activities during such events is essential for ensuring public safety and enabling rapid response from authorities. Traditional monitoring systems often rely on manual surveillance and reporting mechanisms, which may not provide timely situational awareness. This paper presents a Live Event Detection framework for people's safety using Natural Language Processing (NLP) and Deep Learning techniques. The proposed system collects real-time textual data from social media platforms, news feeds, and public communication channels to identify safety-related events and emerging threats. NLP techniques are employed for text preprocessing, sentiment analysis, entity recognition, and feature extraction, while deep learning models are utilized to classify event types and detect critical situations. The framework enables the identification of incidents such as accidents, crowd congestion, violence, emergencies, and public safety concerns in real time. Experimental analysis demonstrates that the proposed approach improves event detection accuracy, enhances situational awareness, and supports proactive safety management. The developed system provides an intelligent and scalable solution for safeguarding people during large-scale public events.