摘要
对有后验概率分布的决策问题进行贝叶斯分析,决策个体对概率的估计偏差未在考虑之内,导致不同决策者的分析结果并无差异,针对该问题,基于等级依赖期望效用模型中概率权重函数影响决策的机制,在贝叶斯分析中引入概率权重函数,体现决策个体对概率的不同估计,能很好克服贝叶斯决策分析的不足.以Prelec概率权重函数为例的分析表明概率权重函数的引入合理有效,可提升决策分析的针对性.
For decision with posterior distribution, if we use Bayesian decision analysis, the deviation of probability is out of consideration, it leads that different individuals' analysis results are the same unreasonably. Based on RDEU's mechanism of probability weight function influences decision, we introduce probability weight function into Bayesian decision analysis. It can reflect the deviation of probability to overcome the shortage of normative Bayesian decision analysis. A case study of Prelec probability weight function shows the introducing is effective and can improve the pertinence of decision analysis.
出处
《数学的实践与认识》
北大核心
2015年第3期148-153,共6页
Mathematics in Practice and Theory
基金
武器装备军内科研基金项目"效用理论在战场目标分配中应用研究"(JNKY2012015)
陆军军官学院科研基金项目"效用理论在战场目标分配中应用研究"(YNKY2012015)
关键词
贝叶斯决策
概率权重函数
等级依赖期望效用模型
Prelec概率权重函数
bayesian decision analysis
probability weight function
the rank-dependentexpected utility model
prelec probability weight function