No label defined (Q4713372)

We propose a logic which is of inductive character and is intended as a tool for data analysis, i.e. our main concern is in discovering dependencies in data and data reduction, which is rather closer to statistical then deductive methods, however to this end we shall use deductive tools. Our main goal is reasoning about knowledge concerning certain reality. We have assumed that knowledge is represented as a value-attribute table, called sometimes information system or knowledge representation system. Representation of knowledge in tabular form, has great advantages in particular for its clarity. It turns out that the data table may be looked at as a set of propositions about the reality and consequently can be treated by means of logical tools, which will be developed. We offer two possibilities here, one based on normal form representation of formulas and the second employing indiscernibility to investigate whether some formulas are true or not. The latter approach, referring to indiscernibility, leads to simple algorithms for data reduction and analysis. Our main concern is the fundamental notion of the decision logic, the decision algorithm, which is a set of decision rules (implications). Formulas can be true or false but the decision algorithm, which is a set of formulas, can not have attributes of truth or falsity. Instead consistency and inconsistency will be the basic features of decision algorithms. In contrast to philosophy of deduction, we stress rather consistency (or inconsistency) of data then their truth (or falsity), and our main interest is not in investigation of theorem proving mechanisms in the introduced logic, but in analysis, in computational terms (decision algorithms, or condition-action rules) of how some facts are derived from data.