Add dataset card
Browse files
README.md
CHANGED
|
@@ -1,36 +1,60 @@
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
- name: category
|
| 17 |
-
dtype: string
|
| 18 |
-
- name: original_sentiment
|
| 19 |
-
dtype: string
|
| 20 |
-
splits:
|
| 21 |
-
- name: validation
|
| 22 |
-
num_bytes: 127753
|
| 23 |
-
num_examples: 20
|
| 24 |
-
- name: test
|
| 25 |
-
num_bytes: 112975
|
| 26 |
-
num_examples: 20
|
| 27 |
-
download_size: 134577
|
| 28 |
-
dataset_size: 240728
|
| 29 |
-
configs:
|
| 30 |
-
- config_name: default
|
| 31 |
-
data_files:
|
| 32 |
-
- split: validation
|
| 33 |
-
path: data/validation-*
|
| 34 |
-
- split: test
|
| 35 |
-
path: data/test-*
|
| 36 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- ar
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
task_categories:
|
| 6 |
+
- multiple-choice
|
| 7 |
+
- text-classification
|
| 8 |
+
pretty_name: Sentiment Analysis MCQ Evaluation Dataset
|
| 9 |
+
tags:
|
| 10 |
+
- sentiment-analysis
|
| 11 |
+
- mcq
|
| 12 |
+
- financial
|
| 13 |
+
- arabic
|
| 14 |
+
- evaluation
|
| 15 |
+
- benchmark
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
---
|
| 17 |
+
|
| 18 |
+
# Sentiment Analysis MCQ Evaluation Dataset
|
| 19 |
+
|
| 20 |
+
Validation and test splits for financial sentiment analysis in MCQ format.
|
| 21 |
+
|
| 22 |
+
## Dataset Structure
|
| 23 |
+
|
| 24 |
+
- **Format**: Multiple choice questions
|
| 25 |
+
- **Language**: Arabic
|
| 26 |
+
- **Domain**: Financial reports
|
| 27 |
+
- **Task**: Sentiment classification
|
| 28 |
+
- **Validation**: 20 examples
|
| 29 |
+
- **Test**: 20 examples
|
| 30 |
+
|
| 31 |
+
## Fields
|
| 32 |
+
|
| 33 |
+
- `id`: Unique identifier
|
| 34 |
+
- `query`: Full MCQ prompt
|
| 35 |
+
- `answer`: Correct answer letter
|
| 36 |
+
- `text`: Question text
|
| 37 |
+
- `choices`: Answer options [a, b, c]
|
| 38 |
+
- `gold`: Correct answer index
|
| 39 |
+
- `category`: Report category
|
| 40 |
+
- `original_sentiment`: Ground truth sentiment
|
| 41 |
+
|
| 42 |
+
## Answer Mapping
|
| 43 |
+
|
| 44 |
+
- a) positive - gold: 0
|
| 45 |
+
- b) negative - gold: 1
|
| 46 |
+
- c) neutral - gold: 2
|
| 47 |
+
|
| 48 |
+
## Usage
|
| 49 |
+
|
| 50 |
+
```python
|
| 51 |
+
from datasets import load_dataset
|
| 52 |
+
|
| 53 |
+
dataset = load_dataset("SahmBenchmark/Sentiment_Analysis_MCQ_eval")
|
| 54 |
+
test_data = dataset['test']
|
| 55 |
+
|
| 56 |
+
for example in test_data:
|
| 57 |
+
print(f"Question: {example['text']}")
|
| 58 |
+
print(f"Choices: {example['choices']}")
|
| 59 |
+
print(f"Correct: {example['answer']} (index: {example['gold']})")
|
| 60 |
+
```
|