ARRINA - Association and Recommendation for Requirements in Natural Language

• Extracts specifications and recommends which are the most adequate for customer requirements
• ARRINA recommends specifications for requirements with an accuracy of 80%
• ARRINA achieves a reduction of testing time (mapping specifications for requirements) of 85%

Input(s):

• Existing Natural Language (NL) Customer Requirements and Design Specifications linked between them
• Input requirements written in NL

Main feature(s):

• Automated extraction and interpretation of requirements and specifications
• Main requirements and specifications
• Representation of the knowledge within a subsystem
• Analysis and correlation of requirements and specifications for a Top-K recommendations

Output(s):

• Top-K recommendations for requirements to specifications
• Knowledge graph representative of the knowledge within a PCS
• Associations rules among requirements and among specifications of a subsystem

• Python 3
• Graphviz
• pandas
• NLTK
• mlxtend
• openpyxl
• xlrd

Requirements engineers, Software architects, Project managers.

Open-source restricted and permission required.