Annotated software requirement corpus

English software requirement text corpus annotated with universal dependencies syntactic and part-of-speech information. Requirements are taken from a variety of domains contained in the open source PURE corpus (accessible at http://nlreqdataset.isti.cnr.it/).

Data available: https://github.com/UniversalDependencies/UD_English-CTeTex/

Input(s):

• N/A

Main feature(s):

• Software and system requirement descriptions annotated in universal dependencies

Output(s):

• N/A

None

Machine learning or natural language processing experts that require training or evaluation data for automatic analysis of software requirements in universal dependencies grammar as a standalone task or as part of a multi-objective task. This corpus can help improve natural language understanding tasks aiming to interpret, validate or analyse software requirements in a requirement validation or verification scenario.

Open source licence CC BY-SA 4.0