Published August 17, 2021
| Version 1.0.0
Dataset
Open
Reproducible Authorship Attribution Benchmark Tasks
Creators
- 1. Indiana University Bloomington
- 2. Duquesne University
Description
Reproducible Authorship Attribution Benchmark Tasks (RAABT) consists of five closed-set authorship identification experiments.
Each task features fixed train and test sets. Four of the five tasks have a test set consisting of writing samples on fixed topics, guaranteeing that test set examples do not overlap with training set examples in terms of subject matter. Data for all tasks is available for download without any restrictions.
The file README.md contains a full description of the data.
Files
raabt-v1.0.0.zip
Files
(3.6 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:7d9b0fc3ffb5d778643f0e24895482ae
|
3.6 MB | Preview Download |
Additional details
Funding
- U.S. National Science Foundation
- SaTC: CORE: Small: Collaborative: Defending Against Authorship Attribution Attacks 1814425