FNFC: 7,060 Software Requirement Statements with 14 Labels

7,060 labeled software requirement statements categorized into 14 classes, created by refining the Fault-prone SRS Dataset from Kaggle. The dataset was developed by MSHD-IAU and last updated on August 10, 2025.

Use Cases

• Train classifiers to distinguish functional from non-functional requirements based on labeled statements.
• Benchmark NLP models for text classification tasks in software engineering.
• Analyze the distribution and characteristics of different requirement types.
• Develop tools for automated requirements analysis and quality assurance.

Strengths

• 7,060 labeled data points provide a substantial corpus for training.
• 14 distinct requirement classes offer a detailed classification scheme.
• High-quality annotations ensured through expert review.
• Dataset is derived from a known Kaggle source, providing a traceable origin.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but specific data formats and file sizes are unknown.
• Freshness should be verified as the last update was August 10, 2025.

Provenance

Source

Refined and re-labeled from the Fault-prone SRS Dataset on Kaggle.

Collection Method

Created by refining and re-labeling an existing dataset, with annotations reviewed by experts.

Freshness

Last updated 2025-08-10 05:33:10