A semantic interpretation of rules in interval sets

An approach to induce classification rules from incomplete information tables is proposed in this paper. A new semantic interpretation of rules in interval sets is given based on the three-way decisions theory. In the framework of three-way decisions theory, an interval set can be divided into three regions: the positive, negative, and boundary regions. In accordance with the three regions, one can make a three-way decision: acceptance, rejection and noncommitment. The three regions enable us to derive three types of decision rules, namely, positive rules for acceptance, negative rules for rejection, and boundary rules for indecision or delayed decision.