Abstract:
One of the problem in data analysis reduces to problems of a vectors subset selection. The NP-completeness of these problems is proved. These problems are connected with searching a vector subset in a given set in Euclidian space under following conditions. The first condition is that the required subset has a given cardinality, and the second one is that this subset includes vectors which are close to each other under the criterion of minimum sum of squared distances. Bibliogr. 13.
Keywords:choice of a vector subset, clustering, algorithmic complexity, NP-completeness.