The approach introduced in this paper depicts the topic of identification and evaluation of stockout consequences, commonly denoted as stockout cost quantification. Our work is motivated by the limited number of approaches dealing with this problem and, primarily in the field of inventory management, a subsequent need for applicable methods providing reliable stockout cost parameters. We focus on the problem of estimating opportunity costs of stockouts as the most difficult cost component to be determined. Therefore, a method to elicit information by confronting relevant decision makers with representative stockout cases (a priori) is presented. Subsequently, a Genetic Programming (GP) approach for learning opportunity cost functions from these case-based decisions is introduced. It is shown on exemplary tests instances that solutions can be generated which converge to structurally similar opportunity cost functions for representative stockout items. Based on a comparison to benchmarks generated by Neural Networks, it can be concluded that the quality of solutions from the GP algorithm is satisfying.