Structure coarsening, entropy and compressed space dimension (Q1433753)

The authors show that it is possible to reduce information systems thanks to reduction algebra. Under some mathematical restrictions the composition of algebras can quantify the entropy of the system. As a consequence, an isomorphism of a discretized space is constructed to obtain a new subspace with random dimension that well characterizes the random aspect of the initial system. As in thermodynamic statistics, the energy is quantified. The results of the authors enable them to quantify a great number of dynamic phenomena such as complex coarsening structure or equilibrium state.