Chapter

Swarm, Evolutionary, and Memetic Computing

Volume 7076 of the series Lecture Notes in Computer Science pp 637-644

The Use of Strategies of Normalized Correlation in the Ant-Based Clustering Algorithm

  • Arkadiusz LewickiAffiliated withUniversity of Information Technology and Management in Rzeszów
  • , Krzysztof PancerzAffiliated withUniversity of Information Technology and Management in Rzeszów
  • , Ryszard TadeusiewiczAffiliated withAGH University of Science and Technology

* Final gross prices may vary according to local VAT.

Get Access

Abstract

The article presents a new approach to the evaluation process associated with the modification of the ant-based clustering algorithm. The main aim of this study is to determine the degree of impact of the proposed changes on the results of the implemented clustering algorithm, whose task is not only to obtain the lowest intra-group variance, but also to self-determine the amount of target classes. These modifications concern both a different way of choosing the radius of perception considering the neighborhood of objects in a search decision space, as well as a use of a completely different metric than the Euclidean one for calculating the dissimilarity of objects based on the components including the normalized angular correlation of objects under consideration.

Keywords

ATTA metaheuristics strategy heuristic decision-making system clustering data data mining unsupervised clustering