Detecting empirical regularities in bases of facts using JSM reasoning

Abstract

The previously introduced definition of empirical regularities (ERs) is considered using JSM reasoning. These definitions used functionals of the inconsistency degree for expandable sets of generated hypotheses. New ER definitions (of empirical laws and empirical trends) are formulated by the predicates of preserving truth-value types (factual truth, factual falsity, factual contradiction, and uncertainty). It is shown that the previously introduced definitions are the consequence of the new definitions. A set of possible ER definition options is considered using various JSM-reasoning strategies (inductive inference rules), which form a distributive lattice. It is proposed to use the projection of the set of JSM strategies on the set of empirical regularities and the inverse mapping of the ER set into the set of JSM strategies. The ER definition is used to amplify K.R. Popper's demarcation criterion, which characterizes a study that uses empirical data as a completed research work.