Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -1,28 +1,60 @@
|
|
1 |
-
---
|
2 |
-
license: cc-by-4.0
|
3 |
-
dataset_info:
|
4 |
-
features:
|
5 |
-
- name: original
|
6 |
-
dtype: string
|
7 |
-
- name: tokens
|
8 |
-
sequence: string
|
9 |
-
- name: labels
|
10 |
-
sequence: string
|
11 |
-
- name: qid
|
12 |
-
sequence: string
|
13 |
-
- name: language
|
14 |
-
dtype: string
|
15 |
-
- name: url
|
16 |
-
dtype: string
|
17 |
-
splits:
|
18 |
-
- name: train
|
19 |
-
num_bytes: 5669993
|
20 |
-
num_examples: 6764
|
21 |
-
download_size: 1906917
|
22 |
-
dataset_size: 5669993
|
23 |
-
configs:
|
24 |
-
- config_name: default
|
25 |
-
data_files:
|
26 |
-
- split: train
|
27 |
-
path: data/train-*
|
28 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-4.0
|
3 |
+
dataset_info:
|
4 |
+
features:
|
5 |
+
- name: original
|
6 |
+
dtype: string
|
7 |
+
- name: tokens
|
8 |
+
sequence: string
|
9 |
+
- name: labels
|
10 |
+
sequence: string
|
11 |
+
- name: qid
|
12 |
+
sequence: string
|
13 |
+
- name: language
|
14 |
+
dtype: string
|
15 |
+
- name: url
|
16 |
+
dtype: string
|
17 |
+
splits:
|
18 |
+
- name: train
|
19 |
+
num_bytes: 5669993
|
20 |
+
num_examples: 6764
|
21 |
+
download_size: 1906917
|
22 |
+
dataset_size: 5669993
|
23 |
+
configs:
|
24 |
+
- config_name: default
|
25 |
+
data_files:
|
26 |
+
- split: train
|
27 |
+
path: data/train-*
|
28 |
+
language:
|
29 |
+
- nl
|
30 |
+
- en
|
31 |
+
- es
|
32 |
+
- pt
|
33 |
+
- el
|
34 |
+
- fr
|
35 |
+
- de
|
36 |
+
pretty_name: winnl
|
37 |
+
---
|
38 |
+
|
39 |
+
# WiNNL
|
40 |
+
|
41 |
+
WikiNews Named entity recognition and Linking (WiNNL) is a multilingual news NER & NEL benchmark based on Wikinews articles.
|
42 |
+
The dataset was created by automatically scraping and tagging news articles, and manually corrected by native speakers to ensure accuracy.
|
43 |
+
|
44 |
+
You can find more information in the paper:
|
45 |
+
https://aclanthology.org/2024.dlnld-1.3.pdf
|
46 |
+
|
47 |
+
The dataset includes the following NER classes in IOB format (`labels`):
|
48 |
+
* **PER** (Person): person names
|
49 |
+
* **LOC** (Location): geographical locations
|
50 |
+
* **ORG** (Organisation): organisations
|
51 |
+
* **AMB** (Ambiguous): entities that had an ambigous wikidata link in the article, and could be classified as multiple NER classes
|
52 |
+
* **DATE** (Date): dates (e.g. "2022-01-01", "5th of January 2022")
|
53 |
+
* **DISEASE** (Disease): diseases (e.g. "cancer", "COVID-19")
|
54 |
+
* **EVT** (Event): events (e.g. "2024 US elections")
|
55 |
+
* **SPE** (Sport Event): sports events (e.g. "World Cup", "Olympics")
|
56 |
+
* **OTH** (Other): other entities that do not fit into any of the above categories
|
57 |
+
|
58 |
+
***Please note that only the PER, ORG and LOC classes have been corrected manually.***
|
59 |
+
|
60 |
+
The `qid` column contains the Wikidata entity identifiers for the entities in the dataset, also in IOB format.
|