Multi-valued logical method for knowledge extraction and updating in real time

Abstract

Familiar methods for knowledge derivation are based on statistical procedures, which are not accompanied by algorithms for updating of knowledge in real time. In this work, the algorithms are implemented successfully using both logical and statistical procedures whereby flexible arrays of experimental data is entered such that older data is removed and newer data is added. These data are packed together in groups and transformed into logical values of functions of multi-valued logic. The functions are accompanied by probability of occurrence of its values, which is evaluated in real time. In this way, a new construction is formed called multi-valued logical probability function (MLPF), which expresses simultaneously two mutually related correspondences-logical and probabilistic. The correspondences are updated in real time. MLPF is a system of production rules, which are updated in real time.