A New Framework of Combinational Evaluation Methodology Based on Statistics

Abstract

Comparing with single evaluation method, combinational evaluation method is more informative, more objective and more accurate. A new framework of combinational evaluation methodology based on statistics is constructed in this paper. At first the Kendall coefficient of consistency is computed to identify whether the different single evaluation methods are compatible. A nonlinear programming model is constructed to determine the weight of different evaluation method result. The effect of the combinational method is determined by the index of MAPE and MSPE. Then 20 listed companies in logistics industry are chosen as sample. Then the evaluation index system is constructed. Based on 4 single evaluation methods which are synthetic index, TOPSIS, efficiency coefficient method and factor analysis, combinational evaluation is constructed with the nonlinear programming model. The exploratory results show that combinational evaluation method is a good choice.