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 1707.07576stat.MLcs.LG
收藏

通过数据集中的属性交互解释分类器

Interpreting Classifiers through Attribute Interactions in Datasets

Finnish Institute of Occupational Health(芬兰职业健康研究所)

Andreas Henelius, Kai Puolamäki, Antti Ukkonen

摘要:

在这项工作中,我们提出了一种新的ASTRID方法,用于调查哪些属性交互分类器在进行预测时利用。分类任务中的属性交互意味着两个或多个属性一起为特定的类标签提供更强的证据。这种相互作用的知识,使模型更容易解释,揭示属性之间的关联。这具有应用,例如,在药物警戒中识别药物之间的相互作用,或在生物信息学中研究单核苷酸多态性之间的关联。我们还展示了如何发现属性分区是相关的数据生成分布的因式分解和经验证明所提出的方法的实用性。

英文摘要:

In this work we present the novel ASTRID method for investigating which attribute interactions classifiers exploit when making predictions. Attribute interactions in classification tasks mean that two or more attributes together provide stronger evidence for a particular class label. Knowledge of such interactions makes models more interpretable by revealing associations between attributes. This has applications, e.g., in pharmacovigilance to identify interactions between drugs or in bioinformatics to investigate associations between single nucleotide polymorphisms. We also show how the found attribute partitioning is related to a factorisation of the data generating distribution and empirically demonstrate the utility of the proposed method.

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。

↑