考虑市场学习行为的贝叶斯决策模型构建及应用
发布时间：2019-01-29 21:38
【摘要】：贝叶斯统计模型对于描述复杂市场环境下的决策行为具有重要意义。当决策者具有根据信息不断更新信念的学习行为时,不能再采用传统的静态稳定参数推断方法进行统计决策。在贝叶斯框架下,可以结合决策主体的主观信念设定相关参数构建模型。基于储蓄-消费模型的研究发现市场学习行为对统计决策的影响显著,构建考虑学习行为的贝叶斯决策模型能够解释过于保守的谨慎行为。通过数值模拟,进一步说明具有学习过程的决策行为更加贴近现实。相关研究结论值得引起对传统风险度量方法的反思。
[Abstract]:Bayesian statistical model is very important to describe the decision behavior in complex market environment. When the decision makers have the learning behavior of updating their beliefs according to the information, they can no longer use the traditional static stable parameter inference method to make statistical decision. Under the Bayesian framework, the model can be constructed by combining the subjective beliefs of decision makers. Based on the economy-consumption model, it is found that market learning behavior has a significant impact on statistical decision-making, and a Bayesian decision model considering learning behavior can explain the conservative cautious behavior. Through numerical simulation, it is further demonstrated that the decision-making behavior with learning process is closer to reality. The relevant research results are worthy of rethinking the traditional risk measurement methods.
【作者单位】： 南昌大学经济管理学院;南昌大学公共管理学院;
【基金】：教育部人文社会科学研究青年基金项目(15YJC630194) 国家自然科学基金青年资助项目(71401069)
【分类号】：C934
本文编号：2417888
[Abstract]:Bayesian statistical model is very important to describe the decision behavior in complex market environment. When the decision makers have the learning behavior of updating their beliefs according to the information, they can no longer use the traditional static stable parameter inference method to make statistical decision. Under the Bayesian framework, the model can be constructed by combining the subjective beliefs of decision makers. Based on the economy-consumption model, it is found that market learning behavior has a significant impact on statistical decision-making, and a Bayesian decision model considering learning behavior can explain the conservative cautious behavior. Through numerical simulation, it is further demonstrated that the decision-making behavior with learning process is closer to reality. The relevant research results are worthy of rethinking the traditional risk measurement methods.
【作者单位】： 南昌大学经济管理学院;南昌大学公共管理学院;
【基金】：教育部人文社会科学研究青年基金项目(15YJC630194) 国家自然科学基金青年资助项目(71401069)
【分类号】：C934
【相似文献】
相关期刊论文 前3条
1 朱金玲;贝叶斯决策分析及改进[J];江苏统计;2000年06期
2 董志强，顾昌耀;贝叶斯决策的影响图分析方法[J];系统工程理论与实践;1994年11期
3 宋翰林;;贝叶斯决策在序贯分析应用[J];现代经济信息;2013年17期
相关会议论文 前1条
1 蒋晓杰;夏蓓;钱存华;;贝叶斯决策分析中有噪声信息的比较[A];第四届全国决策科学/多目标决策研讨会论文集[C];2007年
，本文编号：2417888
本文链接：https://www.wllwen.com/guanlilunwen/lindaojc/2417888.html
