图 书 馆
指数族中具有对数损失水平的独立最优预测Horizon-Independent Optimal Prediction with Log-Loss in Exponential Families |
|
| 课程网址: | http://videolectures.net/colt2013_hedayati_prediction/ |
| 主讲教师: | Fares Hedayati |
| 开课单位: | 加州大学伯克利分校 |
| 开课时间: | 2013-08-09 |
| 课程语种: | 英语 |
| 中文简介: |  |
| 课程简介: | We study online learning under logarithmic loss with regular parametric models. Hedayati and Bartlett (2012) showed that a Bayesian prediction strategy with Jeffreys prior and sequential normalized maximum likelihood (SNML) coincide and are optimal if and only if the latter is exchangeable, which occurs if and only if the optimal strategy can be calculated without knowing the time horizon in advance. They put forward the question what families have exchangeable SNML strategies. We answer this question for one-dimensional exponential families: SNML is exchangeable only for three classes of natural exponential family distributions,namely the Gaussian, the gamma, and the Tweedie exponential family of order 3/2. |
| 关 键 词: | 参数模型; 对数损失; 序列归一化最大似然; 贝叶斯预测策略; 自然指数族 |
| 课程来源: | 视频讲座网公开课 |
| 最后编审: | 2019-05-26：cwx |
| 阅读次数: | 40 |