Linguistic description of adult skeletal age-at-death estimations from fuzzy integral acquired fuzzy sets
Derek T. Anderson, James M. Keller, Melissa F. Anderson, Daniel J. Wescott · 2011
Previously, we introduced a novel method to estimate adult skeletal age-at-death using the Sugeno fuzzy integral (FI). Specifically, we took a multi-hypothesis testing approach to make the classical FI yield a fuzzy set (FS)-valued result, which is not guaranteed to be normal or convex, based on interval-valued sources of information (aging methods). We showed quantitative results for summarizing the FS and comparing the single decision to a known age-at-death. In this article, we extend our prior work and present formulas to measure the uncertainty in the resultant FSs. We generate linguistic descriptions to establish domain standardization for the goal of assisting forensic and biological anthropologists. Specifically, we introduce fuzzy class definitions for age-at-death FSs and we present an OWA contrast approach to measure the degree of specificity in age-at-death FSs.