Aggregation Selection for Hierarchical Fuzzy Signatures: A Comparison of Hierarchical OWA and WRAO

S. Mendis, T. Gedeon.

In general, intelligent decision making systems receive information that is very sparse and which is likely to be hierarchically correlated. Our previous research has show that hierarchical Fuzzy Signatures are effective, efficient, robust and flexible with such inputs. Earlier, we introduced the generalised Weighted Relevance Aggregation Operator (WRAO) for hierarchical Fuzzy Signatures. In this paper, we compare the generalised Ordered Weighted Averaging (GOWA) operator with WRAO to select the best aggregation method for hierarchical Fuzzy Signatures. Additionally, we show a method of learning hierarchical GOWA using the Levenberg-Marquardt optimization Method.

PDF full paper