Conditional classifiers and boosted conditional Gaussian mixture model for novelty detection
作者：
Highlights：
• A new novelty detection classifier is proposed.
• The proposed method is capable of considering the dependencies of relevant random variables.
• No simplifying assumption is made for encoding such dependencies.
• Combining additive modeling and boosting method to learn conditional densities.
• 12%–24% false positive reduction compared to one-class SVM and two other methods.
摘要
•A new novelty detection classifier is proposed.•The proposed method is capable of considering the dependencies of relevant random variables.•No simplifying assumption is made for encoding such dependencies.•Combining additive modeling and boosting method to learn conditional densities.•12%–24% false positive reduction compared to one-class SVM and two other methods.
论文关键词：Novelty detection,Mixture models,Graphical models,Conditional dependence,Conditional density,Additive modeling,Boosting,False positive
论文评审过程：Received 27 June 2017, Revised 4 March 2018, Accepted 23 March 2018, Available online 26 March 2018, Version of Record 24 May 2018.
论文官网地址：https://doi.org/10.1016/j.patcog.2018.03.022
