客观AHP判断矩阵的构造方法研究

Research on the Construction Method of Objective AHP Judgment Matrix

研究目标:为避免AHP赋权受人为主观因素影响,因此对客观AHP赋权方法展开研究。研究方法:从指标相关性、指标相似度和样本空间性三个维度构造判断矩阵,并基于AIC准则给出相对最优方法的遴选原则,通过数值仿真验证文中方法的可行性与实用性,以2018年新疆数字普惠金融指数再合成为例进行对比。研究发现:不同赋权类型会导致样本综合指数的排名发生变化,不同赋权类型下综合指数的平均水平、分位水平、波动情况以及分布形态可能相近也可能存在差异。研究创新:较为系统地给出了客观判断矩阵的构造方法与遴选原则,并基于MATLAB R2016a提供了一套可以调用的标准化代码。研究价值:有效解决了AHP方法中人为主观因素的影响,使决策方案朝着数据驱动方向发展。

Research Objectives:In order to prevent AHP empowerment from being affected by subjective factors,the objective AHP empowerment method is researched.Research Methods:The judgment matrix is constructed from three dimensions of indicator correlation,indicator similarity and sample spatiality.Based on the AIC criterion,the selection principle of the relative best method is given.The feasibility and practicability of the method in this article are verified by numerical simulation.Take the 2018 Xinjiang Digital Financial Inclusive Index re-synthesis as an example for comparison.Research Findings:Different types of empowerment will lead to changes in the ranking of the sample composite index.The average level,quantile level,volatility and distribution pattern of the composite index under different weighting types may be similar or different.Research Innovations:The construction method and selection principle of the objective judgment matrix are systematically given,and a set of standardized codes that can be called is provided based on MATLAB R2016 a.Research Value:Effectively solve the influence of human subjective factors in the AHP method,and make the decision-making plan develop in the data-driven direction.