Incorporating fuzzy operators in the decision network to improve classification reliability

One type of hierarchical fuzzy-operator-based network implementation is investigated. In this approach, we generalized the Dombi operator as an effective component for decision analysis and making. This methodology provides several advantages due to the fact that the input to each node is the evidence supplied by the degree of satisfaction of sub-criteria and the output is the aggregated evidence. Thus, the decision making process is to aggregate and propagate the evidence information through such a hierarchical network. This trainable network is able to perceive and interpret complex decisions by using those transparent fuzzy models. This study examines the behavior of the fuzzy additive operator in more detail and the results show that the proposed framework exhibits reliable decision in the pattern classification domain.