An artificial intelligence-based research and development demand automatic identification and completion method

Abstract

This invention discloses an AI-based method for automatically identifying and completing R&D requirements, comprising the following steps: S1, Requirement parsing: Parsing the requirement document according to predefined tags in the requirement configuration table, adding dimension tags to each functional point; S2, Requirement retrieval and matching: Querying the requirement knowledge base using the tag information after functional point breakdown, combined with a large model and RAG technology, and filtering based on requirement relevance scores; S3, Requirement integration: Comparing the retrieved requirements for differences, and providing completion solutions through the large model; S4, Requirement knowledge base storage: Comparing the requirements from different versions, and vectorizing the necessary labels for the different functional standards, storing them in the knowledge base. This invention can automatically supplement user requirements, improve the completeness of requirement content, and reduce the rework rate and R&D rework costs in requirement implementation.

Images