Learning of parameters of intuitionistic statement networks

The paper presents method of learning weights parameters of intuitionistic statement networks on the base of learning data. The approach is based on the application of substitutional network model which is equivalent to intuitionistic statement networks in the sense of conditional dependency between statements. The substitutional model is a generative model which allows to compute its own parameters even on the base of complete or incomplete learning data. In the paper the general conditions to achieve substitutional model and the method of its analysis are presented. Finally the weights of intuitionistic statement networks are identified on the basis of substitutional model analysis.