On the classification of dynamical data streams using novel “Anti-Bayesian” techniques
作者：
Highlights：
• We present techniques to perform classification in dynamic data streams using novel “Anti-Bayesian” (AB) techniques.
• The novel methodology uses the highly efficient incremental quantile estimators.
• We compare the AB dynamic classification with its Bayesian counterpart.
• The methods are compared for different synthetic and real-life datasets.
• The results show that the AB method, in many cases, clearly outperforms the Bayesian counterpart - which is both surprising and interesting.
摘要
•We present techniques to perform classification in dynamic data streams using novel “Anti-Bayesian” (AB) techniques.•The novel methodology uses the highly efficient incremental quantile estimators.•We compare the AB dynamic classification with its Bayesian counterpart.•The methods are compared for different synthetic and real-life datasets.•The results show that the AB method, in many cases, clearly outperforms the Bayesian counterpart - which is both surprising and interesting.
论文关键词：Anti-Bayesian classification,Data streams,Classification with delay,Incremental quantile estimation
论文评审过程：Received 23 March 2017, Revised 19 October 2017, Accepted 23 October 2017, Available online 3 November 2017, Version of Record 3 November 2017.
论文官网地址：https://doi.org/10.1016/j.patcog.2017.10.031
