punuy-datasets
A collection of quantitative lexical semantic relations datasets, gathered from the academic literature, for the evaluation of semantic measures between concepts or entities. Currently focused on the English and Portuguese languages.
Datasets
English
# | ID | Name | Year | Domain | Sim/Rel | References |
---|
1 | rg65 | Rubenstein and Goodenough (1965) | 1965 | general | sim | paper 📑, website 🌐 |
2 | mervis210 | Mervis 210 | 1975 | general | rel | paper 📑 |
3 | tversky42 | Tversky42 | 1979 | general | sim | paper 📑 |
4 | mc30 | Miller and Charles (1991) | 1991 | general | sim | paper 📑 |
5 | medin34 | Medin 34 | 1993 | general | sim | paper 📑 |
6 | resnik28 | Miller and Charles Replication | 1999 | general | sim | paper 📑, website 🌐 |
7 | rd27 | Resnik and Diab 2000 | 2000 | general | sim | paper 📑, website 🌐 |
8 | ws353 | WordSimilarity-353 | 2002 | general | sim | paper 📑, website 🌐 |
9 | yp130 | YP-130 | 2005 | general | sim | paper 📑 |
10 | bg100k | Princeton Evocation Data 0.4 | 2006 | general | evo | paper 📑, website 🌐 |
11 | mesh2 | MeSH2 | 2006 | biomedical | sim | paper 📑, website 🌐 |
12 | zie55 | Ziegler et al. (2006) | 2006 | general | rel | paper 📑 |
13 | mayoSRS | MayoSRS | 2007 | biomedical | rel | paper 📑, website 🌐 |
14 | miniMSRS | MiniMayoSRS | 2007 | biomedical | rel | paper 📑, website 🌐 |
15 | gm30 | Gracia and Mena (2008) | 2008 | general | rel | paper 📑 |
16 | ps65 | Pirro and Seco (2008) | 2008 | general | sim | paper 📑 |
17 | ws353split | WordSim-353 Similarity and Relatedness Gold Standard | 2009 | general | rel/sim | paper 📑, website 🌐 |
18 | umnsrs | UMNSRS | 2010 | biomedical | rel/sim | paper 📑, website 🌐 |
19 | mt287 | MT287 | 2011 | general | rel | paper 📑, website 🌐 |
20 | atlasify240 | Atlasify240 | 2012 | general | rel | paper 📑, website 🌐 |
21 | mturk771 | MTurk771 | 2012 | general | rel | paper 📑, website 🌐 |
22 | reword26 | REWOrD | 2012 | general | rel | paper 📑, website 🌐 |
23 | scws2003 | SCWS | 2012 | general | sim | paper 📑, website 🌐 |
24 | ma28 | Martinez & Aldana | 2013 | general | sim | paper 📑, website 🌐 |
25 | rel122 | Rel-122 | 2013 | general | rel | paper 📑, website 🌐 |
26 | srw2034 | Stanford Rare Word (RW) Similarity Dataset | 2013 | biomedical | sim | paper 📑, website 🌐 |
27 | wp300 | WP300 | 2013 | general | sim | paper 📑, website 🌐 |
28 | baker143 | Baker 143 | 2014 | general | sim | paper 📑, website 🌐 |
29 | geresid50 | GeReSiD | 2014 | geographical | rel/sim | paper 📑, website 🌐 |
30 | men3000 | MEN | 2014 | general | rel | paper 📑, website 🌐 |
31 | sl7576 | SL7576 | 2014 | general | sim | paper 📑, website 🌐 |
32 | lm659 | Landrigan & Mirman (2015) | 2015 | general | rel/sim | paper 📑, website 🌐 |
33 | simlex999 | SimLex-999 | 2015 | general | sim | paper 📑, website 🌐 |
34 | tr9856 | TR9856 | 2015 | general | rel | paper 📑, website 🌐 |
35 | simverb3500 | SimVerb-3500 | 2016 | general | sim | paper 📑, website 🌐 |
36 | umnsrsMod | UMNSRS (modified) | 2016 | biomedical | rel/sim | paper 📑, website 🌐 |
37 | semeval17 | SemEval-2017 Task 2.1 | 2017 | general | sim | paper 📑, website 🌐 |
38 | ali27 | Ali et al. 2018 | 2018 | general | sim | paper 📑 |
39 | card660 | Card-660 | 2018 | general | sim | paper 📑, website 🌐 |
40 | gtrd | GTRD | 2018 | geographical | rel | paper 📑, website 🌐 |
41 | word19k | Wikipedia Oriented Relatedness Dataset | 2018 | general | rel | paper 📑, website 🌐 |
42 | multisimlex | Multi-SimLex | 2020 | general | sim | paper 📑, website 🌐 |
43 | emocon | Emotion concepts 2022 | 2022 | general | sim | paper 📑, website 🌐 |
44 | sstars13 | Shooting-Stars | 2023 | general | rel | paper 📑, website 🌐 |
Portuguese
Use with Node.js
If you want to programatically use the datasets' data and metadata, install the Node.js wrapper code:
npm install
Then you can access the datasets like this:
import { mc30, rg65 } from "punuy-datasets";
for (const part of mc30.partitions) {
console.log(part.id, part.data.length);
}
Other uses
If you want to use another language or just want to access the data files directly, each dataset has its own folder inside profiles/
.
Each dataset has a dataset.json
file with its metadata and information about its partitions. Then, for each partition with id p_id
, there is a p_id.part.json
file containing the pairs and their ratings.
Adding a new dataset
If you can create the dataset profile yourself, PRs are welcome!
You can find the dataset schema (in Typescript) here.
Or you can open an issue. Please include:
- a link to the dataset homepage (if it exists),
- a link to download the dataset, and
- a link to the paper describing it.
I'll get to it as soon as I have some free time ;)
Types and schema
TypeScript definitions can be found at src/lib/types.ts.
Automatically generated JSON schema definitions can be found at dataset-schema.json and partition-data-schema.json.
Automatically generated documentation can be found at docs/.
Datasets copyright and licenses
Did you find a problem with the licensing of a dataset? We did our best to find and include the licenses to all the datasets and to respect copyrights, but we might have missed something!
Please open an issue and we'll fix it ASAP.
See also
For cross-lingual datasets and datasets in other languages please check the following links:
Bugs and stuff
Open a GitHub issue or, preferably, send me a pull request.
License
The MIT License (MIT)
Copyright (c) 2025 André Santos andrefs@andrefs.com
Permission is hereby granted, free of charge, to any person obtaining a copy of
this software and associated documentation files (the "Software"), to deal in
the Software without restriction, including without limitation the rights to
use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
the Software, and to permit persons to whom the Software is furnished to do so,
subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.