Ordered fuzzy rules generation based on incremental dataset

This paper proposes a novel approach for building transparent knowledge-based systems by generating interpretable fuzzy rules that allow for present dependences between quantitative variables by accounting for uncertainty and the dynamics of their values. In the approach, IF-THEN rules are used to show the conditional relationship between the ordered fuzzy numbers, which contain additional information about the tendencies of variables' value changes. This paper elaborates an approach of mining ordered fuzzy rules from numerical data included in an incremental database. This approach develops the ability to record uncertainty and its change in the context of rapidly changing data. In addition, it is the basis for the development of research on the inference method with ordered fuzzy rules, which may become an indispensable tool for decision-making in an uncertain environment.