Mini-batch bagging and attribute ranking for accurate user authentication in keystroke dynamics

作者：

Highlights：

• We have proposed mini-batch bagging (MINIBAG) method and attribute ranking of one-class naïve Bayes (AR-ONENB) algorithm.

• We have presented attribute-by-attribute data fragmentation technique which is used in MINIBAG method.

• MINIBAG facilitates machine learning algorithms to have an ensemble of multiple models from mini-batches.

• We have introduced a new feature, keystroke index order, based on the typing speed of a user in AR-ONENB algorithm.

• Rate of difference of the rate of mean produces a reliable prediction to the result.

摘要

•We have proposed mini-batch bagging (MINIBAG) method and attribute ranking of one-class naïve Bayes (AR-ONENB) algorithm.•We have presented attribute-by-attribute data fragmentation technique which is used in MINIBAG method.•MINIBAG facilitates machine learning algorithms to have an ensemble of multiple models from mini-batches.•We have introduced a new feature, keystroke index order, based on the typing speed of a user in AR-ONENB algorithm.•Rate of difference of the rate of mean produces a reliable prediction to the result.

论文关键词：Keystroke dynamics,Mini-batch,Bagging,Attribute ranking,One-class naïve Bayes,User authentication

论文评审过程：Received 12 September 2016, Revised 7 February 2017, Accepted 4 May 2017, Available online 5 May 2017, Version of Record 19 May 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.05.002