Information processing in linear vector space

摘要

Strings of information units are written as naive matrices A and linear algebra is used to find quadratical forms ATA and AAT. These are interpreted as position vectors in n-and m -dimensional Euclidean space. Binary information relationships are treated as graphs and bipartite graphs with matrices (A + B) or (A|B). Their quadratical forms GTG and GGT are found and interpreted. Information matrices A and G have twofold symmetry connected with permutations of their rows and columns. The logarithmic measure of this symmetry is entropy. Its limit is the sum of Boltzmann and Shannon entropies.