THIEF DETECTION ALGORITHM FOR ADVANCED SECURITY SYSTEMS BASED ON DYNAMIC NEURAL NETWORKS

COMPUTER SCIENCE & ENGINEERING

This paper explores the development and implementation of a dynamic neural network-based theft detection algorithm tailored for smart security systems. Leveraging the advanced capabilities of neural networks, the proposed algorithm aims to enhance the accuracy and reliability of detecting theft-related activities in various environments. The model is trained and tested using the XGBoost method, demonstrating an overall accuracy of 95%. This high level of performance is reflected in the precise identification of multiple event classes within the security system. For instance, the algorithm achieved near-perfect precision, recall, and F1-scores in most event categories, with Class 1 and Class 3 recording perfect metrics across the board. Despite some variance in performance, particularly in Class 5, the overall results affirm the effectiveness of the dynamic neural network approach in improving theft detection capabilities. The macro and weighted average precision, recall, and F1-scores further support the robustness of the model. This research contributes to the growing field of smart security systems by providing a sophisticated and reliable solution for real-time theft detection, offering significant potential for practical application in enhancing security measures.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.