A statistical approach to the identification of diploid cellular automata based on incomplete observations.

In this paper, the identification problem of diploid cellular automata is considered, in which, based on a series of incomplete observations, the underlying cellular automaton rules and the states of missing cell states are to be uncovered. An algorithm for identifying the rule, based on a statistical parameter estimation method using a normal distribution approximation, is presented. In addition, an algorithm for filling the missing cell states is formulated. The accuracy of these methods is examined in a series of computational experiments.