An on-device gender prediction method for mobile users using representative wordsets

作者：

Highlights：

• We propose an on-device gender prediction method for mobile users.

• The proposed method outperformed existing ones from experiments on realworld data.

• Term popularity and discriminability is important for on-device gender prediction.

摘要

•We propose an on-device gender prediction method for mobile users.•The proposed method outperformed existing ones from experiments on realworld data.•Term popularity and discriminability is important for on-device gender prediction.

论文关键词：Gender prediction,Mobile text data,Representative wordsets,Word evaluation measures,On-device analytics

论文评审过程：Received 17 April 2016, Revised 21 July 2016, Accepted 1 August 2016, Available online 3 August 2016, Version of Record 9 August 2016.

论文官网地址：https://doi.org/10.1016/j.eswa.2016.08.002