constructing multiattribute utility functions for portfolio choice in shipping companies’ stocks

Random multi-attribute utility function allows encoding objects and preferences, which helps the investor score the outcome and select an optimal portfolio choice; however, these models disregard the dependency between the alternative values in different attributes. this study develops a model that can account for the dependency in calculations. for this purpose, the concept of conditional probability was used to develop the model. in this process, the value of each alternative in each criterion was calculated concerning the probability of occurrence of different values in the other criteria; therefore, all possible dependency states were taken into account in calculating the marginal utility value of each alternative. the model was applied to the data from two shipping companies taken from the singapore stock exchange, and the results show the consistency of the findings with the reference set results, which proves the applicability and validity of the proposed model.