Huber损失最小化学习法

最新推荐文章于 2022-11-02 11:09:22 发布

止于至玄

最新推荐文章于 2022-11-02 11:09:22 发布

阅读量1.8k

点赞数

分类专栏： Machine Learning 文章标签：机器学习

本文链接：https://blog.csdn.net/philthinker/article/details/68947986

版权

Huber回归结合了ℓ1损失和ℓ2损失，以平衡模型的鲁棒性和准确性。尽管ℓ1损失可能导致异常值对结果有较大影响，但为了解决这个问题，Tukey回归可以作为一个值得考虑的替代方案。

摘要由CSDN通过智能技术生成

Huber regression
In least square learning methods, we make use of $\ell_{2}$ loss to make sure that we get a suitable outcome. However, in the robust point of view, it is always better to make use of the least absolute as the main criterion, i.e.

θ^L A = arg min θ J L A (θ), J L A (θ) = \sum i = 1 n | r i |

$\hat{\theta}_{LA}=\arg\min_{\theta}J_{LA}(\theta),\quad J_{LA}(\theta)=\sum_{i=1}^{n}|r_{i}|$
where

ri=fθ(xi)−yi $r_{i}=f_{\theta}(x_{i})-y_{i}$ is the residual error. By doing so, it is possible to make the learning method more robust at the cost of accuracy.
In order to balance robustness and accuracy, Huber loss may be a good alternative:

ρ H u b e r (r) = {r 2 / 2 η | r | - η 2 / 2 (

最低0.47元/天解锁文章

止于至玄

关注

0
点赞
踩
4

收藏

觉得还不错? 一键收藏
0
评论
Huber损失最小化学习法

Huber regression In least square learning methods, we make use of ℓ2\ell_{2} loss to make sure that we get a suitable outcome. However, in the robust point of view, it is always better to make use of
复制链接

扫一扫