Description
Programmatic Theme: Data Science
Abstract: Lexical searches for medication names in Twitter suffer from low recall due to misspellings or ambiguity with common words. We present Kusuri, an Ensemble Learning classifier able to identify tweets mentioning drug products and dietary supplements. On a corpus made of all tweets posted by 112 Twitter users, with only 0.26% mentioning medications, Kusuri obtained an F1-score of 78.8%, a score high enough to ensure it is ready to be integrated in public health pipelines.
Learning Objective: - constructing a corpus of tweets with an equal number of tweets mentioning drugs and tweets not mentioning drugs
- building and parametrizing an ensemble of recurrent neural networks to detect tweets mentioning drug names
- evaluating the ensemble of neural networks
Authors:
Davy Weissenbacher (Presenter)
University of Pennsylvania
Abeed Sarker, Emory University
Ari Klein, University of Pennsylvania
Karen O’Connor, University of Pennsylvania
Arjun Magge Ranganatha, Arizona State University
Graciela Gonzalez-Hernandez, University of Pennsylvania