Elicitation of Decisionmaker Preference By Artificial Neural Networks

The classical elicitation methods are not robust when decisionmaker distort or misperceive probabilities. So, which makes it difficult for using the methods in certain applications. This paper presents a new model to elicit the decisionmaker preferences by artificial neural networks (ANNs). This model simulating human thought and cognition is more consistent with the real utility of a decisionmaker. A BP neural network of 3-layers was designed to elicit a decisionmaker utility, and the result was superior to classical elicitation method (i.e. CE). In a word, ANNs present a new method and insight for us to solve the utility elicitation question