Hybrid Algorithm for Inductive Cluster Analysis Using Density Distribution Data Evaluation

2015;
: pp. 178 - 190
Автори: 
Lurie Irina, Osipenko Volodymyr, Lytvynenko Volodymyr, Taif Muchamed Ali, Kornilovska Natalia

Kherson National Technical University. Kherson, Berislavske Shosse 24, Ukraina,
National University of Life and Environmental Sciences of Ukraine, Kyiv, pr. Heroiv Oborony 15, E-mail: immun56@gmail.com

In this article proposed a new clustering technique, which is based on two methods: density algorithm DBSCAN and inductive objective clustering algorithm. Experimentally proved that the combination of two these methods can solve the problem of recognition of clusters of different nonlinear form, and greatly increase the accuracy in the detection of complex objects.