Semantic enrichment of decision rules: a framework for improving formal decision contexts

Yanling Wang, Liwei Sha, Li Zou, Hengfei Li, Luis Martínez, Chris Nugent, J. Liu

Abstract: A Formal Decision Context (FDC) is a knowledge representation for decision-making, in which decision implications capture the dependencies between conditions and conclusions, primarily based on non-semantic information. Lack of semantic information can lead to rule redundancy and a lack of corresponding background knowledge in FDC. This paper addresses the limitations of FDC with regard to the lack of semantic information by introducing a semantic formal decision context (SFDC) that integrates domain ontologies into description logic. Hence, we construct an SFDC with semantic knowledge that enriches the decision-making process. We propose a semantic decision implication framework based on SFDC that leverages the reasoning capabilities of description logics to enhance the comprehensiveness and reduce redundancy in decision implication. To more effectively demonstrate that our method significantly reduces information redundancy while preserving the completeness of semantic decision implications, we have validated the algorithm for generating the simplest semantic decision implication set across six publicly available datasets. The experimental results demonstrate that our algorithm achieves remarkable performance in improving the reduction efficiency of decision implications.

External IDs:doi:10.1016/j.ins.2025.122330