Large Matrices Inversion Using the Basis Exchange Algorithm
Leon Bobrowski *
Faculty of Computer Science, Białystok University of Technology, Poland and Institute of Biocybernetics and Biomedical Engineering, PAS, Warsaw, Poland.
*Author to whom correspondence should be addressed.
Abstract
Data exploration tasks often require inversion of large matrices. The paper presents a new method of matrices inversion, which uses the basis exchange algorithm controlled by the convex and piecewise linear (CPL) inversion criterion function. Using basis exchange algorithms might increase the dimension of the inverted matrices and computational efficiency of the inversion tasks. Basis exchange algorithms are based on the Gauss-Jordan transformation which is used e.g. in the famous Simplex algorithm applied in linear programming.
Keywords: Data exploration, large matrices inversion, basis exchange algorithm, Gauss-Jordan transformation, convex and piecewise linear (CPL) criterion functions.