orestxherija / smm4h2018

Code for my submission to the Shared Task of the 2018 workshop on Social Media Mining for Health Applications

Social Media Mining for Health Applications Shared Task

Will be updated soon with the code. Unfortunately, we are not permitted to release the datasets, but they will soon appear in a Codalab competition where people can try their models' performance. The model weights have been uploaded.

The relevant paper can be found here. Below you can find the .bib entry for the paper.

@InProceedings{W18-5910,
  author = 	"Xherija, Orest",
  title = 	"Classification of Medication-Related Tweets Using Stacked Bidirectional LSTMs with Context-Aware Attention",
  booktitle = 	"Proceedings of the 2018 EMNLP Workshop SMM4H: The 3rd Social Media Mining for Health Applications Workshop and Shared Task",
  year = 	"2018",
  publisher = 	"Association for Computational Linguistics",
  pages = 	"38--42",
  location = 	"Brussels, Belgium",
  url = 	"http://aclweb.org/anthology/W18-5910"
}

Sub-task 1 Leaderboard (11 participants)

Distinguish tweets that mention names of medications or dietary supplements from those that do not. The definitions of drugs and dietary supplements is taken from the FDA.

	Precision	Recall	F1
THU_NGN	0.933	0.904	0.918
UChicagoCompLx	0.937	0.891	0.914
IRISA	0.922	0.906	0.914
Tub-Oslo	0.917	0.907	0.912
CIC-NLP	0.920	0.899	0.910
UZH	0.927	0.878	0.902
Techno	0.905	0.855	0.879
IIT_KGP	0.918	0.840	0.877
LILU	0.841	0.860	0.850
ART	0.785	0.880	0.830
ClaC	0.788	0.769	0.778

Sub-task 2 Leaderboard (8 participants)

Distinguish tweets that mention personal medication intake, possible medication intake or no intake.

	Precision	Recall	F1
UChicagoCompLx	0.654	0.783	0.713
Light	0.492	0.467	0.479
Tub-Oslo	0.464	0.466	0.465
IRISA	0.434	0.501	0.465
IIT_KGP	0.408	0.407	0.408
UZH	0.371	0.437	0.401
CLaC	0.402	0.366	0.383
Techno	0.327	0.432	0.372

Sub-task 3 Leaderboard (9 participants)

Distinguish tweets that contain mentions of adverse drug reaction those that do not.

	Precision	Recall	F1
THU_NGN	0.442	0.636	0.522
IRISA	0.378	0.649	0.478
UZH	0.455	0.436	0.445
Tub-Oslo	0.638	0.317	0.424
ART	0.332	0.547	0.413
UChicagoCompLx	0.370	0.464	0.411
CIC-NLP	0.314	0.529	0.394
Techno	0.434	0.344	0.383
IIT_KGP	0.189	0.643	0.292

Sub-task 4 Leaderboard (9 participants)

Distinguish tweets that mention behavior related to influenza vaccination from those that do not. Data annotators labeled tweets to answer the binary question Does this message indicate that someone received, or intended to receive, a flu vaccine?

	Precision	Recall	F1
CARRDS	0.918	0.859	0.887
Techno	0.870	0.859	0.865
Light	0.824	0.897	0.859
Tub-Oslo	0.840	0.872	0.855
UChicagoCompLx	0.791	0.923	0.852
IRISA	0.867	0.833	0.850
LILU	0.829	0.808	0.818
ClaC	0.700	0.897	0.787
IIT_KGP	0.800	0.769	0.784