Abstract
The Density-Based Spatial Clustering of Applications with Noise (DBSCAN) is a well-known data-mining method capable of localization of accident blackspots of the road network based on the already existing road accident database records. However, its parameterization raises many problems, as its operation is significantly different from the traditional Sliding Window (SW) method. This paper presents a Particle Swarm Optimization (PSO) based method to find a base parameter set for the DBSCAN method which gives similar results to the already existing SW. The fitness function of the PSO algorithm is based on the similarity of accident blackspots, which needs a definition of a novel metric. The evaluation results show that the DBSCAN method used with the recommended parameter set is capable to give similar results to the SW method used by road safety experts.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
