Abstract: The grouping information problem manifests itself in two main forms. Namely these are: recovering function, represented by empirical data (observations) and problem of classification (clusterization). In both forms the choice of fundamental” representatives” are of principal importance: arguments – function characteristics for the first case and ‘feature vector” in the second one case. Vector variants of representatives selection, namely choice of , are practiced as a rule. This choice is determined by availability of a highly developed technique for processing of objects in Euclidean spaces. This technique includes, particularly, SVD and Moore-Penrose? inversion for the matrixes of linear operators between Euclidean spaces of -type. Realization of selection step or case by the standard recurrent procedures is represented in the article as well as development of processing technique for Euclidean space of type (Euclidean space all matrixes of fixed dimension). It is turn out that selection procedure for in both cases can be designed on the base of so called neurofunctional transformations (NfT-transformations). As to SVD and Moore-Penrose? technique are proposed in this case.

Keywords: Feature vectors, information aggregating, generalized artificial neuronets, vector corteges, matrix corteges, linear operator between cortege spaces, Single Valued Decomposition for cortege linear operators.

ACM Classification Keywords: G.2.m. Discrete mathematics: miscellaneous, G.2.1 Combinatorics. G.3 Probability and statistics, G.1.6. Numerical analysis I.5.1.Pattern Recognition H.1.m. Models and Principles: miscellaneous.

Link:

‘FEATURE VECTORS’ IN GROUPING INFORMATION PROBLEM IN APPLIED MATHEMATICS: VECTORS AND MATRIXES

Donchenko V., Zinko T., Skotarenko F.

http://foibg.com/ibs_isc/ibs-28/ibs-28-p13.pdf