Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
json
Sub-tasks:
sentiment-classification
Languages:
Czech
Size:
10K - 100K
License:
fewshot-goes-multilingual-admin
commited on
Commit
•
eed49f9
1
Parent(s):
6f5f882
Update README.md
Browse files
README.md
CHANGED
@@ -20,3 +20,32 @@ task_categories:
|
|
20 |
task_ids:
|
21 |
- sentiment-classification
|
22 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
task_ids:
|
21 |
- sentiment-classification
|
22 |
---
|
23 |
+
|
24 |
+
|
25 |
+
# Dataset Card for Mall.cz Product Reviews (Czech)
|
26 |
+
|
27 |
+
|
28 |
+
## Dataset Description
|
29 |
+
|
30 |
+
The dataset contains user reviews from Czech eshop <mall.cz>
|
31 |
+
Each review contains text, sentiment (positive/negative/neutral), and automatically-detected language (mostly Czech, occasionaly Slovak) using [lingua-py](https://github.com/pemistahl/lingua-py)
|
32 |
+
The dataset has in total (train+validation+test) 30,000 reviews. The data is balanced.
|
33 |
+
|
34 |
+
Train set has 8000 positive, 8000 neutral and 8000 negative reviews.
|
35 |
+
Validation and test set each have 1000 positive, 1000 neutral and 1000 negative reviews.
|
36 |
+
|
37 |
+
|
38 |
+
## Dataset Features
|
39 |
+
|
40 |
+
Each sample contains:
|
41 |
+
- `review_id`: unique string identifier of the review.
|
42 |
+
- `rating_str`: string representation of the rating - "pozitivní" / "neutrální" / "negativní"
|
43 |
+
- `rating_int`: integer representation of the rating (1=positive, 0=neutral, -1=negative)
|
44 |
+
- `comment_language`: language of the review (mostly "cs", occasionaly "sk")
|
45 |
+
- `comment`: the string of the review
|
46 |
+
|
47 |
+
|
48 |
+
## Dataset Source
|
49 |
+
|
50 |
+
The data is a processed adaptation of [Mall CZ corpus](https://liks.fav.zcu.cz/sentiment/).
|
51 |
+
The adaptation is label-balanced and adds automatically-detected language
|