Skip to content
/ ELSA Public

A dataset for Entity-Level Sentiment Analysis

License

Notifications You must be signed in to change notification settings

ltgoslo/ELSA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ELSA - a Norwegian Entity-level sentiment dataset

The enclosed files contain our Entilty-level annotations of the Norec_fine dataset. Our annotations provide separate annotations for each volitional entity in each document. The Norec_fine texts are drawn from NoReC. The NoReC repo contains metadata about the texts in our dataset, e.g. category, date of publishing, and author(s).

The paper presenting this dataset can be found here.

doc_data

This export contains the different texts with their reference ID back to NoReC, a list of the volitional entity IDs found in the text, the document text. split into sentences, the character-level sentence boundaries, and the train-dev-test where this document belongs, according to Norec_fine. 'norec_id', 'elsa_ids', 'sentence_texts', 'sentence_boundaries', 'split'.
This information is used by the other exports, to retrieve the document text, and each sentence separately, accessed by the 'sentence_texts' list indices.

elsa_overall

This exports contain one entry per volitional entity and presents the annotators' evaluation of the overall sentiment that this text conveys regarding the given entity. 'norec_id', 'elsa_id', 'split', 'text', 'elsa_overall_sentiment'

elsa_sentences

This export contains the aggregated sentiment in each sentence regarding the given volitional entity. Sentences with no sentiment annotation are regarded "Neutral" 'norec_id', 'elsa_id', 'split', 'sentence_idx', 'sentencelevel_agg'

elsa_lowlevel

This export contains each sentiment annotation in the dataset, apart from the overall sentiment presented in elsa_overall.
The annotations here is the basis for the aggregations presented in elsa_sentences. Each reference category is available for analysis, and one may wish to aggregate this information according to a different heuristics.
'reference_cat', 'sentiment', 'begin', 'end', 'text', sentence_idx', 'norec_id', 'elsa_id', 'split'

Contains data especially related to our paper at the LAW-XVIII 2024 workshop.

About

A dataset for Entity-Level Sentiment Analysis

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published