A study on stock ranking and selection strategy based on UTA method under the condition of inconsistence

In this paper, a strategy for ranking and selecting stocks based on UTA(UTilitès Additives) model is presented. In the process of training sample selection, two algorithms: Random Selection and Stepwise Elimination are adopted to improve the performance of the UTA-based method. With the information of objective data of stocks combined with the DM(Decision Maker)´s subjective preference statements, a utility value function can be induced to reflect the DM´s preference using the UTA-based algorithm through preference disaggregation. Thus, a ranking on the whole set of stocks will be derived from the utility value function, which could be different with the objective ranking. The inconsistencies between the derived ranking and the ranking of stocks according to the objective data reveal possible overestimation or underestimation of stocks and then provide aid for the DM in the process of selecting stocks. This strategy can be also used interactively by the DM. At last, a numerical experiment on a 40-stock dataset selected from China´s stock market has been done to illustrate the method presented in this paper.