Risk-Sensitive Online Learning
Risk-Sensitive Online Learning
复制标题
风险敏感的在线学习
DOI：
10.1007/11894841_18
复制
发表时间：
2006
期刊：
影响因子：
3
通讯作者：
Jennifer Wortman Vaughan
中科院分区：
文献类型：
--
作者：
Eyal Even;Michael Kearns;Jennifer Wortman Vaughan
我们在考虑在线学习的问题时，不仅要与最好的专家或股票的回报竞争，还要在回报和风险之间进行最佳权衡。受金融应用的启发，我们考虑了平衡收益和风险的两种常用度量：夏普比率[9]和马科维茨均值方差准则[8]。我们首先提供负面结果，确定在这些措施下不后悔算法的不可能性，从而与仅返回设置形成鲜明对比。然后，我们证明了Cesa-Bianchi等人的最新算法在改进的双准则风险-收益度量下实现了非平凡性能，并给出了一个改进的最佳专家算法，该算法对于“局部”版本的均值-方差准则实现了无遗憾。我们在最近六年的标准普尔500指数数据集上对传统的在线算法和新的风险敏感算法进行了实验比较，发现改进的最佳专家算法在夏普比率、MV和累积财富方面优于传统算法。据我们所知，本文提出了在最坏情况在线学习标准模型中明确风险考虑的研究。
We consider the problem of online learning in settings in which we want to compete not simply with the rewards of the best expert or stock, but with the best trade-off between rewards and risk. Motivated by finance applications, we consider two common measures balancing returns and risk: the Sharpe ratio [9] and the mean-variance criterion of Markowitz [8]. We first provide negative results establishing the impossibility of no-regret algorithms under these measures, thus providing a stark contrast with the returns-only setting. We then show that the recent algorithm of Cesa-Bianchi et al. [5] achieves nontrivial performance under a modified bicriteria risk-return measure, and give a modified best expert algorithm that achieves no regret for a “localized” version of the mean-variance criterion. We perform experimental comparisons of traditional online algorithms and the new risk-sensitive algorithms on a recent six-year S&P 500 data set and find that the modified best expert algorithm outperforms the traditional with respect to Sharpe ratio, MV, and accumulated wealth. To our knowledge this paper initiates the investigation of explicit risk considerations in the standard models of worst-case online learning.