Comparison and Analysis of Novel Score-Variance Portfolio Models based on Methods for Ranking Fuzzy Numbers.

Ranking the fuzzy numbers plays an important role in fuzzy decision-making, but it is challenging to rank multiple fuzzy numbers comprehensively and effectively. In previous studies, various properties and factors of fuzzy numbers are considered to construct the ranking scoring formula. Therefore, in theory, the uncertainty of fuzzy numbers can be better defuzzified, and the formula can measure the characteristics of fuzzy numbers. In application, it is meaningful and valuable to investigate the use of fuzzy number's ranking score in fuzzy portfolio to measure characteristics of the asset. Based on the classical mean-variance (M-V) model and four kinds of ranking score formulas, this paper constructs the score-variance (S-V) models. Finally, a numerical experiment is presented to illustrate the feasibility and validity of our proposed S-V model. Compared with the M-V model, on one hand the results of our proposed models are generally consistent with the classic M-V model; on the other hand, the return of our proposed model is slightly smaller in some part, but it performs better in skewness and Sharpe Ratio. [ABSTRACT FROM AUTHOR]