| 关于贝叶斯估计的进一步讨论 |
| |
| 引用本文： | 陶靖轩.关于贝叶斯估计的进一步讨论[J].中国计量学院学报,2001,12(3):11-15. |
| |
| 作者姓名： | 陶靖轩 |
| |
| 作者单位： | 中国计量学院管理科学与工程系,浙江,杭州,310034 |
| |
| 摘 要： | 文章则从绝对差损失函数与相对差损失函数出发 ,先导出两种新的贝叶斯解——中值估计和积分比估计 ;然后再对实际问题进行具体应用 ;在选定同类先验分布条件下将以上结果与传统的条件期望估计、最大后验估计相比较 ,探讨了各类贝叶斯解的优良性 .
|
| 关 键 词： | 共轭分布 核函数 中值估计 积分比估计 条件期望估计 最大后验估计 |
| 文章编号： | 1004-1540(2001)03-0011-05 |
| 修稿时间： | 2001年3月20日 |
Further discuss on the bayesian estimate methods |
| |
| TAO Jing\|xuan.Further discuss on the bayesian estimate methods[J].Journal of China Jiliang University,2001,12(3):11-15. |
| |
| Authors: | TAO Jing\|xuan |
| |
| Abstract: | In this paper, two new Bayesian solutions-estimations of ratio integration and of central value are obtained. They come from the loss functions of the absolute error and the relative error respectively. Then some new applications are discussed. Under some conditions, estimate of central valne has its robustuess superiority over the estimate of conditional expectation or estimate of posterior probability. |
| |
| Keywords: | Cojgugate distribution Kernel function estimate of central Value estimate of conditional expectation estimate of ratio integration estimate of Posterior probability |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
