A link-bridged topic model for cross-domain document classification

作者：

Highlights：

• We propose a Link-Bridged Topic model for cross-domain document classification.

• LBT utilizes an auxiliary link network to discover the co-citation relationship.

• LBT combines the content information and link structures into a graphical model.

• LBT outperforms both multi-view learning and single-view transfer baselines.

摘要

•We propose a Link-Bridged Topic model for cross-domain document classification.•LBT utilizes an auxiliary link network to discover the co-citation relationship.•LBT combines the content information and link structures into a graphical model.•LBT outperforms both multi-view learning and single-view transfer baselines.

论文关键词：Cross-domain,Document classification,Transfer learning,Auxiliary link network

论文评审过程：Received 11 January 2012, Revised 11 February 2013, Accepted 14 May 2013, Available online 22 June 2013.

论文官网地址：https://doi.org/10.1016/j.ipm.2013.05.002