Abstract
In this paper, we compare the four evaluation methods to construct the multiattribute utility function based on the data of decision maker's preferences (A_??_A', A∼A', A_??_A). The first method is the Keeney's approach constructing an additive or multiplicative utility function based on the expected utility hypothesis. The second is the Probit analysis approach which estimates the weights of additive utility by maximum likelihood estimation method. The third is the Bayesian approach which estimates the weights of additive utility by Bayesian estimation method. The last is the discrimination function approach which estimates the weights of additive utility by the statistical correlation ratio maximization method. These methods are applied to the property value analysis.
