Improved DBSCAN Radar Signal Sorting Algorithm Based on Rough Set

摘要

Spatial clustering algorithm based on density in noise environment (DBSCAN algorithm) is a classical density clustering algorithm. In view of the traditional DBSCAN clustering algorithm parameters set the unreliability of depend on human experience, this paper introduced variable precision rough set theory, the parameters of weighted clustering, which can effectively solve the similarity between radar signal data to consider the problem of inadequate, combining with the inverse trigonometric function after data preprocessing methods. The original distance could automatically obtain convenient neighborhood parameter, improve the reliability of DBSCAN algorithm, effectively improve the separation accuracy improved DBSCAN clustering algorithm. Simulation results verify the effectiveness of the proposed model.