# FNFC: Functional & Non-Functional Requirements Classification Dataset

**Link:** [Hugging Face – FNFC Dataset](https://huggingface.co/datasets/Mashhad-Azad-University/FNFC-Functional_Non-Functional_Classification)

## Overview
The FNFC dataset is a labeled collection of **7,060 software requirement statements** categorized into **14 requirement classes**, designed for research and modeling in requirements classification. It was created by refining and re-labeling the **Fault-prone SRS Dataset** from [Kaggle](https://www.kaggle.com), ensuring high-quality annotations through expert review.

## Data Collection & Labeling
The labeling process was carried out by **five professionals** with expertise in software project leadership, systems analysis, or software requirements engineering:

- **2 experts** with over **10 years** of experience  
- **3 experts** with **2–3 years** of experience in requirements engineering  

Each expert received:
- A **questionnaire**  
- Relevant **documentation**  
- Clear labeling **criteria** for identifying functional and non-functional requirements  

Experts were given **20 days** to complete labeling at their own pace. Each record was assigned to one of the following classes:

### Classes
- **F** – Functional Requirements  
- **A** – Availability  
- **AU** – Autonomy  
- **FT** – Fault Tolerance  
- **LF** – Look and Feel  
- **LL** – Legal & Licensing  
- **M** – Maintainability  
- **O** – Inter-Operability  
- **P** – Portability  
- **PE** – Performance  
- **R** – Reliability  
- **SC** – Scalability  
- **SE** – Security  
- **US** – Usability  

## Dataset Structure
| Field       | Description |
|-------------|-------------|
| `id`        | Unique record identifier |
| `requirement` | Requirement statement text |
| `label`     | One of the 14 defined classes |



## Purpose
The dataset provides a robust benchmark for:
- Machine learning models for requirements classification  
- Natural language processing experiments in software engineering  
- Studies comparing classification methods for functional vs. non-functional requirements  

## Access
The dataset is publicly available:  
[Hugging Face – FNFC Dataset](https://huggingface.co/datasets/Mashhad-Azad-University/FNFC-Functional_Non-Functional_Classification)

## Citation
If you use this dataset in academic research, please cite:

```bibtex
@dataset{fnfc_dataset_2025,
  title={FNFC: Functional & Non-Functional Requirements Classification Dataset},
  author={Mashhad Azad University},
  year={2025},
  url={https://huggingface.co/datasets/Mashhad-Azad-University/FNFC-Functional_Non-Functional_Classification}
}

## Contact
Created by Mahdi Kabootari & Younes Abdeahad

📬 kabootarimahdi2@gmail.com
📬 abdeahad.y3@gmail.com
📬 e.kheirkhah@gmail.com

```