Impact of Anonymization on Sentiment Analysis of Twitter Postings

Lampoltshammer, Thomas J.; Thurnay, Lörinc; Eibl, Gregor

doi:10.1007/978-3-658-27495-5_5

Cited by 4 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The three papers mentioned above conclude that de-identification does not have a (strong) negative effect on the model performance regarding downstream NLP tasks. Finally, although not clinical text, Lampoltshammer et al (2019) show that anonymization can cause significant negative changes in the sentiment analysis performance on Twitter data. This work, however, goes beyond existing related work, as we conduct the first analysis regarding the anonymization of clinical text and the effects thereof on ML models.…”

Section: Related Workmentioning

confidence: 88%

Clinical Text Anonymization, its Influence on Downstream NLP Tasks and the Risk of Re-Identification

Ben Cheikh Larbi,

Burchardt,

Roller

2023

Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: Student Research W

View full text Add to dashboard Cite

While text-based medical applications have become increasingly prominent, access to clinical data remains a major concern. To resolve this issue, further de-identification and anonymization of the data are required. This might, however, alter the contextual information within the clinical texts and therefore influence the learning and performance of possible language models. This paper systematically analyses the potential effects of various anonymization techniques on the performance of state-of-theart machine learning models based on several datasets corresponding to five different NLP tasks. On this basis, we derive insightful findings and recommendations concerning text anonymization with regard to the performance of machine learning models. In addition, we present a simple re-identification attack applied to the anonymized text data, which can break the anonymization.

show abstract

Section: Related Workmentioning

confidence: 88%