High dimensional pattern recognition using the recursive hyperspheric classification algorithm

The Recursive Hyperspheric Classification (RHC) algorithm is a novel technique that excels in classifying multivariate, labeled datasets, which may be used for identification of unknown feature vectors. When training the classifier system, RHC meticulously dissects an n-dimensional space into a taxonomic structure of classifiers, or hyperspheres. This algorithm methodically partitions the space into labeled classes. Structure and order materialize from this constant, recursive process of spawning hyperspheres; this constructs an organized hierarchical tree that, when traversed, allows labels, or classes, to be inferred from the current knowledgebase. In benchmarking, RHC boasts superior results compared to modern classification techniques. This paper offers a comprehensive examination of the RHC algorithm, including various improvements to the original version of the algorithm as well as new results of its application.