Transition interval estimation to elicit membership functions in fuzzy evaluation models of animal production systems

The problem of eliciting membership functions for developing expert-knowledge based fuzzy models to evaluate animal production systems is studied. Three existing elicitation methods are considered and a classification of these elicitation methods is given based on the experts' assessment and the response type during the elicitation process. Transition interval estimation is proposed as a new elicitation method that completes our classification. Properties of the proposed method are discussed and are compared to three other elicitation methods from the literature. Guidelines are given on how to select an elicitation method suitable for a given purpose. An illustrative example is given regarding the welfare of laying hens in an egg production system.