Swarm intelligence for self-organized clustering

Algorithms implementing populations of agents which interact with one another and sense their environment may exhibit emergent behavior such as self-organization swarm intelligence. Here a system, called Databionic (DBS), is introduced able to adapt itself structures high-dimensional data characterized by distance and/or density-based in the space. By exploiting interrelations intelligence, emergence, DBS serves an alternative approach optimization global objective function task clustering. The omits usage parameter-free because it searches for Nash equilibrium during its annealing process. To our knowledge, first combining these approaches. Its clustering can outperform common methods K-means, PAM, single linkage, spectral clustering, model-based Ward, if no prior knowledge about available. A central problem correct estimation number clusters. This addressed visualization topographic map allows assessing It known that all algorithms construct clusters, irrespective set contains clusters or not. In contrast most other algorithms, identifies, meaningless (natural) performance demonstrated on benchmark data, are constructed pose difficult problems two real-world applications.