arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2212.01897cs.LG
收藏

分类和回归问题中实例硬度的表征

Characterizing instance hardness in classification and regression problems

Universidade Federal de São Paulo (UNIFESP)(圣保罗联邦大学);Instituto Tecnológico da Aeronáutica (ITA)(航空技术学院)

Gustavo P. Torquette, Victor S. Nunes, Pedro Y. A. Paiva, Lourenço B. C. Neto, Ana C. Lorena

摘要:

最近在机器学习(ML)文献中的一些工作已经证明了评估哪些观察结果最难准确预测其标签的有用性。通过识别这些实例,可以检查它们是否有任何应该解决的质量问题。还可以根据观察的难度水平设计学习策略。本文提出了一组元特征,旨在表征数据集的哪些实例最难准确预测其标签,以及为什么会这样,即实例硬度度量。分类和回归问题都被考虑。构建和分析具有不同复杂程度的合成数据集。还提供了一个包含所有实现的Python包。

英文摘要:

Some recent pieces of work in the Machine Learning (ML) literature have demonstrated the usefulness of assessing which observations are hardest to have their label predicted accurately. By identifying such instances, one may inspect whether they have any quality issues that should be addressed. Learning strategies based on the difficulty level of the observations can also be devised. This paper presents a set of meta-features that aim at characterizing which instances of a dataset are hardest to have their label predicted accurately and why they are so, aka instance hardness measures. Both classification and regression problems are considered. Synthetic datasets with different levels of complexity are built and analyzed. A Python package containing all implementations is also provided.

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。

↑