Data-Informed Fuzzy Measures for Fuzzy Integration of Intervals and Fuzzy Numbers

摘要: The fuzzy integral (FI) with respect to a measure (FM) is powerful means of aggregating information. most popular FIs are the Choquet and Sugeno, research focuses on these two variants. arena FM much more populated, including numerically derived FMs such as Sugeno $\lambda$ -measure decomposable measure, expert-defined FMs, data-informed FMs. drawback that one must know something about relative values input sources. However, there many problems where this information unavailable, crowdsourcing. This paper or those computed by an algorithm analyzes some property data itself, gleaning importance each source they provide. original instantiation agreement FM, which assigns high confidence combinations sources agree another. extends upon our previous work in proposing uniqueness additive for interval-valued evidence. We then extend number (FN)-valued inputs. demonstrate proposed interval FN evidence both synthetic real-world data.