On Some Possibilities of Using AI Methods in the Search for Cause-And-Effect Relationships in Accumulated Empirical Data

The possibilities of using artificial intelligence methods to solve the tasks of making responsible decisions based on Big Data analysis in a strictly limited time are discussed. An approach is proposed in which forecasting the development of a situation and decision-making uses mathematical models of an interpolation-extrapolation type, where empirical dependencies of a causal type serve as a “tool” for describing patterns initially hidden in constantly accumulating empirical data. The proposed approach is based on the heuristic of causal similarity. A variant of the mathematical formalization of this heuristic by algebraic means is presented, which is based on clarifying the concept of similarity as a binary algebraic operation, leading to the well–known mathematical technique of Galois closures and fix points as a platform for data analysis. The mathematical features of the proposed formalization are discussed, as well as some significant advantages and limitations of such an intelligent data analysis scheme. The presented mathematical “toolkit” description is added by examples of the application of the developed approach in modern high–tech medicine - in the analysis and prediction of outcome of neurosurgical operations of human brain tumors.