One of the trends of mechanical product design is intelligent, which needs Expert System with knowledge processed capability to collect requirements of user quickly and efficiently through artificial intelligence technology, analyze user's requirements and convert the requirements into information of product concepts design which is used to design products that meet customer's need.This article applies the Natural Language Understanding technology which is core topic and technique of language and textual processing in the artificial intelligence area on the mechanical product design, focuses on studying the phenomenon and the rcovery strategy of omissible noun. Targeted the omitted term in subject and object, which take up 82% in the noun ellipsis, we analyze them from syntax, semantics and discourse level in NLU and propose the recovery framework of general omitting type and recovery strategy of six special types based on the statistical data of experimental reasearch on corpus, and use the theme of discourse to verified the recovery result, thus, design the common model of the recovery of omissible noun.Finally, embed the recovery module we designed into Expert System, apply which on modern mechanical product design and achieve good understanding result, the research can support virtualization and intelligence of product design in future. |