Algorithms of compacting the knowledge base by sample data (Q1355326): Difference between revisions

One of the pressing problems that arise in constructing a knowledge base \(\Omega\) in some field is the extraction of knowledge from empirical samples. Here the term ``knowledge base'' stands for a set of inference rules of the type IF-THEN. Such knowledge bases are usually called one-step. The knowledge in one-step-bases is presented as logical statements in the form of a conjunction of causes. In \textit{G. S. Lbov} [Methods of processing different-type experimental data (1981; Zbl 0449.62001)], an algorithm (called the GUHA method) was suggested for constructing and verifying those conjunctive statements. However, in that paper there was no mention of any method of constructing a minimal set of statements required to describe the empirical table observed. In the present paper, we suggest some optimal algorithms to solve this problem.