|
|
--- |
|
|
license: mit |
|
|
--- |
|
|
# Argument Scheme Classification Dataset |
|
|
|
|
|
This dataset contains annotated arguments labeled with one of **7 argumentation schemes**. It is intended for training and evaluating argument classification models. |
|
|
|
|
|
## π§ Task |
|
|
|
|
|
The goal is to classify a given **argument** into one of the following **7 argumentation schemes**: |
|
|
|
|
|
- **AV**: Argument from Values |
|
|
- **PR**: Argument from Practical Reasoning |
|
|
- **AC**: Argument from Consequences |
|
|
- **CO**: Argument from Commitments |
|
|
- **AH**: Ad Hominem |
|
|
- **VV**: Victimization |
|
|
- **AS**: Argument from Sign |
|
|
|
|
|
## π Dataset Structure |
|
|
|
|
|
The dataset is split into two parts: |
|
|
|
|
|
- `train.csv` β 700 labeled examples |
|
|
- `validation.csv` β 175 labeled examples |
|
|
|
|
|
Each file contains: |
|
|
|
|
|
| Column | Description | |
|
|
|-----------|---------------------------------| |
|
|
| Argument | The input argument as free text | |
|
|
| Label | One of the 7 argument schemes | |
|
|
|
|
|
## π Format Example |
|
|
|
|
|
```csv |
|
|
Argument,Label |
|
|
"The decree I signed, which converts environmental fines into investments, was signed by me...",AC |
|
|
"Responsible tourism brings with it awareness of the ecosystem and promotes sustainability.",AV |
|
|
|