← Search

Data · dataset · 2022

Dataset for Benchmarking Aspect Level Sentiment Classification on Reviews Requiring Coreference Resolution

Listed in Borealis

We present the ALSC-CR dataset for Aspect Level Sentiment Classification (ALSC).

Description

This dataset can be used to benchmark ALSC model performance on reviews that require coreference resolution (CR) ability to perform well. For example, CR is needed to predict the sentiment for "food" in the sentence - "I had food, it was bad." Here, "food" has a negative sentiment only because it is referring to the same entity as "it" which in turn is associated with a negative sentiment.

Links

Where it is published

Catalogue records · 1

Topics

Provenance · 1 source records, 7 field assertions
SourceKeyLast seenRaw
Borealisdoi:10.5683/SP3/HSKJEY13 d agoJSON v1
FieldAssertionExtractorEvidence
concepts[field].dataverse_subject:computer-and-information-sciencesource · borealisdata caconnector:borealisdata_ca@1.0.0/subjects
created_datesource · borealisdata caconnector:borealisdata_ca@1.0.0
descriptionsource · borealisdata caconnector:borealisdata_ca@1.0.0/description
publication_datesource · borealisdata caconnector:borealisdata_ca@1.0.0
titlesource · borealisdata caconnector:borealisdata_ca@1.0.0/name
updated_datesource · borealisdata caconnector:borealisdata_ca@1.0.0
version_labelsource · borealisdata caconnector:borealisdata_ca@1.0.0