One-Class Classification based on searching for the problem features limits

作者：

Highlights：

• We introduce and analyze a new version of our novelty detection method, namely FBDOCC.

• We introduce a heuristic aimed at reducing the training time of the method.

• We perform a rigorous theoretical study of the algorithm complexity.

• We carry out a number of experiments in order to compare the proposed methods to the methods in the state of the art.

摘要

•We introduce and analyze a new version of our novelty detection method, namely FBDOCC.•We introduce a heuristic aimed at reducing the training time of the method.•We perform a rigorous theoretical study of the algorithm complexity.•We carry out a number of experiments in order to compare the proposed methods to the methods in the state of the art.

论文关键词：One-Class Classification,Anomaly detection,Novelty detection,Nearest neighbor rule

论文评审过程：Available online 3 June 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.05.037