Approximating Joint Probability Distributions Given Partial Information

被引:9
|
作者
Montiel, Luis V. [1 ]
Bickel, J. Eric [1 ]
机构
[1] Univ Texas Austin, Grad Program Operat Res & Ind Engn, Austin, TX 78712 USA
基金
美国国家科学基金会;
关键词
maximum entropy; incomplete information; probabilistic dependence; analytic center; stochastic optimization; dynamic programming; sequential exploration; practice; DECISION-MAKING;
D O I
10.1287/deca.1120.0261
中图分类号
C93 [管理学];
学科分类号
12 ; 1201 ; 1202 ; 120202 ;
摘要
In this paper, we propose new methods to approximate probability distributions that are incompletely specified. We compare these methods to the use of maximum entropy and quantify the accuracy of all methods within the context of an illustrative example. We show that within the context of our example, the methods we propose are more accurate than existing methods.
引用
收藏
页码:26 / 41
页数:16
相关论文
共 50 条