Health issue identification in social media based on multi-task hierarchical neural networks with topic attention

作者：

Highlights：

• Hierarchical neural networks with topic attention are proposed to identify health issue in posts.

• Health issue identification task and domain category classification task are learned parallelly in the multi-task mechanism.

• Topic attention weight is adopted in text encoder.

• The proposed approach outperforms several deep neural networks models.

摘要

•Hierarchical neural networks with topic attention are proposed to identify health issue in posts.•Health issue identification task and domain category classification task are learned parallelly in the multi-task mechanism.•Topic attention weight is adopted in text encoder.•The proposed approach outperforms several deep neural networks models.

论文关键词：Multi-task learning,Topic attention

论文评审过程：Received 29 June 2020, Revised 19 May 2021, Accepted 21 May 2021, Available online 31 May 2021, Version of Record 11 June 2021.

论文官网地址：https://doi.org/10.1016/j.artmed.2021.102119