Regularized Kernel Algorithms for Support Estimation

Frontiers in Applied Mathematics and Statistics

Rudi, Alessandro⁴  Odone, Francesca⁵  De Vito, Ernesto⁶  Verri, Alessandro⁷ 

[1] di Genova, Italy;DIBRIS, UniversitàDipartimento di Matematica, UniversitàINRIA—Sierra team—ÉLaboratory for Computational and Statistical Learning, Istituto Italiano di Tecnologia, Italy;cole Normale Supérieure, France

关键词: Support estimation;    Kernel PCA;    novelty detection;    Dimensionality reduction.;    Regularized kernel methods;   

DOI  :  10.3389/fams.2017.00023

学科分类：数学（综合）

来源: Frontiers

【 摘 要 】

In the framework of non-parametric support estimation, we study the statistical properties of an estimator defined by means of Kernel Principal Component Analysis (KPCA). In the context of anomaly/novelty detection the algorithm was first introduced by Hoffmann in 2007. We also extend to above analysis to a larger class of set estimators defined in terms of a filter function.

【 授权许可】

CC BY