NLP@VCU: Identifying Adverse Effects in English Tweets for Unbalanced Data

Darshini Mahendran, Cora Lewis, Bridget McInnes


Abstract
This paper describes our participation in the Social Media Mining for Health Application (SMM4H 2020) Challenge Track 2 for identifying tweets containing Adverse Effects (AEs). Our system uses Convolutional Neural Networks. We explore downsampling, oversampling, and adjusting the class weights to account for the imbalanced nature of the dataset. Our results showed downsampling outperformed oversampling and adjusting the class weights on the test set however all three obtained similar results on the development set.
Anthology ID:
2020.smm4h-1.29
Volume:
Proceedings of the Fifth Social Media Mining for Health Applications Workshop & Shared Task
Month:
December
Year:
2020
Address:
Barcelona, Spain (Online)
Venue:
SMM4H
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
158–160
Language:
URL:
https://aclanthology.org/2020.smm4h-1.29
DOI:
Bibkey:
Cite (ACL):
Darshini Mahendran, Cora Lewis, and Bridget McInnes. 2020. NLP@VCU: Identifying Adverse Effects in English Tweets for Unbalanced Data. In Proceedings of the Fifth Social Media Mining for Health Applications Workshop & Shared Task, pages 158–160, Barcelona, Spain (Online). Association for Computational Linguistics.
Cite (Informal):
NLP@VCU: Identifying Adverse Effects in English Tweets for Unbalanced Data (Mahendran et al., SMM4H 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/paclic-22-ingestion/2020.smm4h-1.29.pdf
Code
 nlpatvcu/smm4h