Solution of a double clusterization problem with the use of self- organization. (English. Russian original) Zbl 0728.68103
Sov. J. Autom. Inf. Sci. 21, No. 3, 77-82 (1988); translation from Avtomatika 1988, No. 3, 74-79 (1988).
Summary: An algorithm for solving a double clusterization problem is considered. The criterion used in the algorithm is that of minimum center distances, which is analogous to the consistency criterion. An example shows that, under certain conditions, the algorithm proposed here can be used to choose informative parameters for a regression model.
MSC:
68T10 | Pattern recognition, speech recognition |
62H30 | Classification and discrimination; cluster analysis (statistical aspects) |