The EduRABSA Dataset and the ASQE-DPT Annotation Tool
Authors/Creators
-
1.
University of Auckland
Description
This repository contains the full dataset and data annotation tool from our work: EduRABSA: An Education Review Dataset for Aspect-based Sentiment Analysis Tasks
The full information and instructions are available at https://github.com/yhua219/edurabsa_dataset_and_annotation_tool
The EduRABSA Dataset
The EduRABSA dataset is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) License.
Education Review ABSA (EduRABSA) is a manually annotated student review text dataset for multiple Aspect-based Sentiment Analysis tasks, including:
- Aspect-(opinion-category)-Sentiment Quadruplet Extraction (ASQE)
- Aspect-(opinion)-Sentiment Triplet Extraction (ASTE)
- Aspect Sentiment Classification (ASC; a.k.a. Aspect Polarity Classification; APC)
- Aspect Category Detection (ACD) / Aspect-opinion Categorisation (AOC)
- Aspect-Opinion Pair-Extraction (AOPE)
- Aspect Extraction (AE)
- Opinion Extraction (OE)
The dataset consists of 6,500 pieces of stratified samples of public tertiary student review text in the English language released in 2020-2023 on courses ("course review", N=3,000), teaching staff ("teacher review", N=3,000), and university ("university review", N=500).
Dataset Information
Please visit https://github.com/yhua219/edurabsa_dataset_and_annotation_tool
Unannotated dataset Source
| Review Type | Dataset Name | Publish Year | Licence | Total Entries | Sampled (N=6,500) |
|---|---|---|---|---|---|
| Course review | Course Reviews University of Waterloo [1] | October 2022 | CC0: Public Domain | 14,810 | 3,000 |
| Teacher review | Big Data Set from RateMyProfessor.com for Professors' Teaching Evaluation [2] | March 2020 | CC BY 4.0 | 19,145 | 3,000 |
| University review | University of Exeter Reviews [3] | June 2023 | CC0: Public Domain | 557 | 500 |
[1]: Waterloo Course Reviews. Course Reviews University of Waterloo. October 2022.
[2]: RateMyProfessor Dataset. Big Data Set from RateMyProfessor.com for Professors' Teaching Evaluation. March 2020.
[3]: Exeter Reviews. University of Exeter Reviews. June 2023.
The ASQE-DPT Annotation Tool
The ASQE-DPT data annotation tool is licensed under a MIT License.
ASQE-DPT is a manual ABSA annotation tool that we extended based on the ABSA Dataset Prepare Tool (DPT) (source) for more comprehensive and challenging ABSA tasks.
ASQE_DPT is a no-code, no-installation small HTML file that can be used locally and offline to protect data security/privacy.
The .zip file contains the annotation tool and real unannotated and annotated data samples.
For usage instructions, please visit https://github.com/yhua219/edurabsa_dataset_and_annotation_tool
Files
ASQE-DPT.zip
Files
(31.7 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:aa97a8fdf63b1c4f15e47a2cac3baa78
|
524.8 kB | Preview Download |
|
md5:f2f8b534e0ea83b6f3f7915237542b56
|
31.2 MB | Preview Download |