464-475
MATRICES OF INDICATORS OF RECOVERABLE KNOWLEDGE.
Authors: S. U. Zhanatauov ; R.B. Seitkamzina
Number of views: 161
The problem is solved in the article: for the given 5 matrices of eigenvectors of different dimensions (they were calculated using real data [2-6]), find 2 matrices A + qp, B + pp of eigenvectors containing all indicators of 5 matrices of eigenvectors Cnn [2- 6]. Without changing indicators, transform the solution to this problem into the solution from [1], where the input object is the diagonal matrix ?pp = diag (?1, ..., ?p), ?1> ...> ?p> 0, ?1 + ... + ?p = p, p + q = n, p? q. it is required to find the values ??of the elements of 2 model submatrices Zmq, Zmp of the matrix Zmn = [Zmq/Zmp], consisting of m values ??of n z-variables, n = q + p, q?p. The set of z-variables is divided into 2 groups: the first group combines the z-variables z1, ..., z5, the second z6, ..., z9.) :( 1 / m) UTU = ? (1) pp, (1 / m) VTV = ? (2) pp, (1 / m) UTV = ?2pp = diag (?21, ..., ?2p), ?21> ...> ?2p> 0. Model matrices A + qp and B + pp must have algebraic properties of orthonormal matrices corresponding to positive eigenvalues.