Saved in:
Bibliographic Details
Main Authors: Baumartz, Daniel, Bagci, Mevlüt, Henlein, Alexander, Konca, Maxim, Lücking, Andy, Mehler, Alexander
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.14626
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912076267520000
author Baumartz, Daniel
Bagci, Mevlüt
Henlein, Alexander
Konca, Maxim
Lücking, Andy
Mehler, Alexander
author_facet Baumartz, Daniel
Bagci, Mevlüt
Henlein, Alexander
Konca, Maxim
Lücking, Andy
Mehler, Alexander
contents If sentiment analysis tools were valid classifiers, one would expect them to provide comparable results for sentiment classification on different kinds of corpora and for different languages. In line with results of previous studies we show that sentiment analysis tools disagree on the same dataset. Going beyond previous studies we show that the sentiment tool used for sentiment annotation can even be predicted from its outcome, revealing an algorithmic bias of sentiment analysis. Based on Twitter, Wikipedia and different news corpora from the English, German and French languages, our classifiers separate sentiment tools with an averaged F1-score of 0.89 (for the English corpora). We therefore warn against taking sentiment annotations as face value and argue for the need of more and systematic NLP evaluation studies.
format Preprint
id arxiv_https___arxiv_org_abs_2410_14626
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle You Shall Know a Tool by the Traces it Leaves: The Predictability of Sentiment Analysis Tools
Baumartz, Daniel
Bagci, Mevlüt
Henlein, Alexander
Konca, Maxim
Lücking, Andy
Mehler, Alexander
Computation and Language
If sentiment analysis tools were valid classifiers, one would expect them to provide comparable results for sentiment classification on different kinds of corpora and for different languages. In line with results of previous studies we show that sentiment analysis tools disagree on the same dataset. Going beyond previous studies we show that the sentiment tool used for sentiment annotation can even be predicted from its outcome, revealing an algorithmic bias of sentiment analysis. Based on Twitter, Wikipedia and different news corpora from the English, German and French languages, our classifiers separate sentiment tools with an averaged F1-score of 0.89 (for the English corpora). We therefore warn against taking sentiment annotations as face value and argue for the need of more and systematic NLP evaluation studies.
title You Shall Know a Tool by the Traces it Leaves: The Predictability of Sentiment Analysis Tools
topic Computation and Language
url https://arxiv.org/abs/2410.14626