Datasets:
Add quality score (1-10) to articles
Browse files- README.md +27 -1
- data/train.jsonl +2 -2
README.md
CHANGED
|
@@ -46,7 +46,8 @@ UVW 2026 (Underthesea Vietnamese Wikipedia) is a cleaned and processed dataset o
|
|
| 46 |
"title": "Việt Nam",
|
| 47 |
"content": "Việt Nam, tên chính thức là Cộng hòa Xã hội chủ nghĩa Việt Nam...",
|
| 48 |
"num_chars": 45000,
|
| 49 |
-
"num_sentences": 500
|
|
|
|
| 50 |
}
|
| 51 |
```
|
| 52 |
|
|
@@ -57,6 +58,31 @@ UVW 2026 (Underthesea Vietnamese Wikipedia) is a cleaned and processed dataset o
|
|
| 57 |
- `content` (string): Cleaned article content
|
| 58 |
- `num_chars` (int): Number of characters in content
|
| 59 |
- `num_sentences` (int): Estimated number of sentences
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 60 |
|
| 61 |
### Data Splits
|
| 62 |
|
|
|
|
| 46 |
"title": "Việt Nam",
|
| 47 |
"content": "Việt Nam, tên chính thức là Cộng hòa Xã hội chủ nghĩa Việt Nam...",
|
| 48 |
"num_chars": 45000,
|
| 49 |
+
"num_sentences": 500,
|
| 50 |
+
"quality": 8
|
| 51 |
}
|
| 52 |
```
|
| 53 |
|
|
|
|
| 58 |
- `content` (string): Cleaned article content
|
| 59 |
- `num_chars` (int): Number of characters in content
|
| 60 |
- `num_sentences` (int): Estimated number of sentences
|
| 61 |
+
- `quality` (int): Quality score from 1-10 based on article metrics
|
| 62 |
+
|
| 63 |
+
### Quality Score
|
| 64 |
+
|
| 65 |
+
Quality score (1-10) is computed based on Wikipedia quality research:
|
| 66 |
+
|
| 67 |
+
- **Length score (40%)**: Article comprehensiveness based on character count
|
| 68 |
+
- **Sentence score (30%)**: Content depth based on number of sentences
|
| 69 |
+
- **Density score (30%)**: Readability based on average sentence length
|
| 70 |
+
|
| 71 |
+
| Score | Count | Percentage |
|
| 72 |
+
|-------|-------|------------|
|
| 73 |
+
| 2 | 124,937 | 11.2% |
|
| 74 |
+
| 3 | 496,169 | 44.4% |
|
| 75 |
+
| 4 | 224,161 | 20.0% |
|
| 76 |
+
| 5 | 111,906 | 10.0% |
|
| 77 |
+
| 6 | 121,956 | 10.9% |
|
| 78 |
+
| 7 | 25,178 | 2.3% |
|
| 79 |
+
| 8 | 10,905 | 1.0% |
|
| 80 |
+
| 9 | 2,958 | 0.3% |
|
| 81 |
+
| 10 | 54 | 0.0% |
|
| 82 |
+
|
| 83 |
+
References:
|
| 84 |
+
- [Wikipedia Language-Agnostic Quality](https://meta.wikimedia.org/wiki/Research:Prioritization_of_Wikipedia_Articles/Language-Agnostic_Quality)
|
| 85 |
+
- [Automatic Quality Assessment of Wikipedia Articles](https://dl.acm.org/doi/10.1145/3625286)
|
| 86 |
|
| 87 |
### Data Splits
|
| 88 |
|
data/train.jsonl
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d884303d6c3cde2832f8d5b56b045729538faff55d68627871f1b3fb21113ca9
|
| 3 |
+
size 1795288217
|