@MISC{_c.gpuswarm, author = {}, title = {C.GPU Swarm intelligence Data clustering}, year = {}}
Flocking implementation to be run on the NVIDIA GPU. Performance gains ranged from 30 to 60 times improvement of the GPU over the 3GHz CPU implementation. © 2012 Elsevier B.V. All rights reserved. 1. Problem statement and background Cluster analysis is a descriptive data mining task, which involves dividing a set of data objects into a number of groups, called clusters. The motivation behind clustering a set of data is to find its inherent structure and expose that structure as a set of groups [1]. The data objects within each group should exhibit a large degree of similarity while the similarity among different clusters should be minimal [2]. The need for fast, efficient data analysis has driven the research community to continually develop and improve data clustering methods. One method, multiple species flocking clustering [3], a nature-inspired computational model for simulating the dynamics of flocks of entities, is used for high dimensional unstructured data clustering. This method takes an agent-based approach and relies on emergent organization to effectively cluster data. The effectiveness of this approach relies on the organization that arises through a group of agents interacting through simple rules. In the case of data clustering, similar data sets flock together, loosely organizing themselves according to subject. This method hasmetwith success in clustering highdimensional datasets better than traditional methods such as K-means [3]. Unfortunately the method is highly computational intensive and requires hours of computational time to generate acceptable results when analyzing