Pages: 189-198
Published: 30.10.2017
Abstract: The article formulated and solved the optimization problem of modeling and developed an algorithm for calculating the missing non-dominant elements of the spectrum of the correlation matrix. Selected from IM PCA [1,4,8,12] of its submodels, problems, integers, real numbers, multidimensional objects-matrices of values of z- and y-variables, their matrices of values of pair correlations, variances, values of f-parameters of the spectrum (y-variable variance matrices) serve as initial data for the development and solution of an optimization problem with an objective function for the newly considered b-variables (positive and in magnitude less than 1) and their functions of limitations. An algorithm for modeling the missing elements of the spectrum of the correlation matrix was realized and tested on real data [7]: (f1, f2, f4) =>(b2, ...,bn) =>(1, ..., , ..., n).
Key words: missing elements of the spectrum, 6 f-parameters of the spectrum.
|