OPTIMIZING K-MEANS AND DBSCAN CLUSTERING ALGORITHMS FOR SMART CITY TRAFFIC MANAGEMENT

Abstract

This work investigates the optimization of K-Means and DBSCAN clustering algorithms for Smart City traffic management.
As urban traffic systems become increasingly complex, effective clustering techniques are crucial for analyzing and
managing traffic patterns. The research compares the performance of K-Means, a centroid-based algorithm, and DBSCAN, a
density-based approach, using real-world traffic datasets. Experimental results demonstrate that while K-Means offers
quicker execution and efficiency in handling large datasets, it struggles with accurately clustering complex traffic patterns.
Conversely, DBSCAN shows superior clustering quality and higher accuracy in detecting anomalies and noise, albeit at the
cost of increased computational time and memory usage. The findings suggest that a hybrid approach leveraging the
strengths of both algorithms could provide a more robust solution for traffic management in Smart Cities.