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 1909.04060cs.LGastro-ph.IMcs.AIstat.COstat.ML
收藏

一种灵活的异常检测框架

A Flexible Framework for Anomaly Detection via Dimensionality Reduction

Institute for Research in Fundamental Sciences (IPM)(基础科学研究所(IPM));University of Geneva(日内瓦大学);African Institute for Mathematical Sciences(非洲数学科学研究所);South African Radio Astronomy Observatory(南非射电天文台);South African Astronomical Observatory(南非天文台);University of Cape Town(开普敦大学)

Alireza Vafaei Sadr, Bruce A. Bassett, Martin Kunz

摘要:

异常检测具有挑战性,特别是对于高维的大型数据集。在这里,我们探讨了一个通用的异常检测框架的基础上降维和无监督聚类。我们发布了DRAMA,一个通用的Python包,它实现了具有广泛内置选项的通用框架。我们测试了各种各样的模拟和真实的数据集,在多达3000个维度的DRAMA,并发现它的鲁棒性和高度竞争力与常用的异常检测算法,特别是在高维度。DRAMA框架的灵活性允许在一些异常示例可用时进行显著优化,使其成为在线异常检测、主动学习和高度不平衡数据集的理想选择。

英文摘要:

Anomaly detection is challenging, especially for large datasets in high dimensions. Here we explore a general anomaly detection framework based on dimensionality reduction and unsupervised clustering. We release DRAMA, a general python package that implements the general framework with a wide range of built-in options. We test DRAMA on a wide variety of simulated and real datasets, in up to 3000 dimensions, and find it robust and highly competitive with commonly-used anomaly detection algorithms, especially in high dimensions. The flexibility of the DRAMA framework allows for significant optimization once some examples of anomalies are available, making it ideal for online anomaly detection, active learning and highly unbalanced datasets.

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。

↑