AN ADAPTIVE MACHINE LEARNING FRAMEWORK FOR PERSONALIZED WOMEN'S SAFETY

Abstract

Women's safety has become a major social concern due to the increasing number of crimes, including harassment, assault, kidnapping, and domestic violence. Although several safety applications have been developed, most existing systems primarily focus on emergency response after an incident occurs and lack intelligent mechanisms forpredicting potential risks or providing proactive safety assistance. To address these limitations, this paper presents An Adaptive Machine Learning Framework for Personalized Women's Safety Using Crime Prediction, Emergency Alert, and Intelligent Route Navigation. The proposed framework integrates machine learning techniques with location-based services to provide a comprehensive and personalized safety solution. A Neural Network model is trained using a historical women crime dataset to predict the expected number of crimes across different geographical regions, enabling users to identify crime-prone areas before planning their travel. The system incorporates multiple functional modules, including secure user registration and authentication, crime prediction, intelligent route navigation to nearby police stations, emergency panic alerts, and crime heatmap visualization. During emergency situations, users can activate the panic module to send an immediate email notification containing their current location to a registered emergency contact. Additionally, the navigation module assists users in locating nearby police stations, while the heatmap module provides a graphical representation of crime intensity to improve situational awareness. The framework is implemented using Python, Django, MySQL, TensorFlow, and Google Maps API, providing an efficient, scalable, and userfriendly web application. Experimental evaluation demonstrates that the proposed system effectively predicts crime patterns and integrates predictive analytics with emergency communication and locationbased assistance. The proposed framework contributes to enhancing women's safety by combining preventive and reactive safety measures within a single intelligent platform, thereby supporting informed decision-making and rapid emergency response.