Interpretability and Explainability of LSP Evaluation Criteria

Logic Scoring of Preference (LSP) is a soft computing decision method for evaluation and selection of complex objects and alternatives using logic criteria. LSP criteria are based on graded logic aggregation structures that in most cases have a canonical form of tree. Such trees aggregate degrees of truth or degrees of fuzzy membership. The aggregators are graded logic functions. Each node in the tree of logic aggregators has a specific semantic identity - it has the interpretation, role, meaning, and importance for the decision maker. The semantic identity of all arguments can be used to develop explainable LSP criteria, and to provide explanation of all results generated in the process of evaluation, comparison, and selection of complex objects and alternatives. In this paper we propose explainability parameters and use them in decision making to provide the explainability of both the results of evaluation of a single object, and the results of comparison and selection of multiple competitive alternatives.