arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2112.06101stat.MLcs.LG
收藏

随机森林泛化误差的置信区间

Confidence intervals for the random forest generalization error

Insper Institute of Education and Research(Insper教育与研究学院)

Paulo C. Marques F

AI总结:

该研究提出利用随机森林标准训练的副产品直接计算泛化误差置信区间的方法,无需数据拆分与重训练,兼具低计算成本与良好覆盖率和合适宽度收缩率的优势。

摘要:

我们证明,随机森林标准训练过程的副产品不仅能提供众所周知且几乎无计算成本的模型泛化误差袋外点估计,还能为计算泛化误差的置信区间提供直接路径,无需数据拆分和模型重训练过程。除构建的计算成本低之外,模拟实验表明这些置信区间具有良好的覆盖率,且其宽度随训练样本量呈现合适的收缩速率。

英文摘要:

We show that the byproducts of the standard training process of a random forest yield not only the well known and almost computationally free out-of-bag point estimate of the model generalization error, but also give a direct path to compute confidence intervals for the generalization error which avoids processes of data splitting and model retraining. Besides the low computational cost involved in their construction, these confidence intervals are shown through simulations to have good coverage and appropriate shrinking rate of their width in terms of the training sample size.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑