Deep learning model for classifying drug abuse risk behavior in tweets

Document Type

Conference Proceeding

Publication Date

7-24-2018

Abstract

Social media such as Twitter can provide urgently needed drug abuse intelligence to support the campaign of fighting against the national drug abuse crisis. We employed a targeted tweet collection approach and a two-staged annotation strategy that combines conventional annotation with crowdsourced annotation to produce annotated training dataset. In this demo, we share deep learning models trained in a boosting manner using the data from the two-staged annotation method and unlabeled data collection to detect drug abuse risk behavior in tweets.

Identifier

85051129449 (Scopus)

ISBN

[9781538653777]

Publication Title

Proceedings 2018 IEEE International Conference on Healthcare Informatics Ichi 2018

External Full Text Location

https://doi.org/10.1109/ICHI.2018.00066

First Page

386

Last Page

387

This document is currently not available here.

Share

COinS