Analysis of HE(***) in patent Chinese-English machine translation

In machine translation, Ηe(***) mainly acts as noun, preposition and conjunction for patent text. Different parts of speech have different ways to translate in machine translation system, especially when it acts as conjunction which refers to coordinate structure. In traditional linguistic, there are papers researched Ηe(***) but only concerned on how to distinguish the preposition and conjunction of it. In papers that studied coordinate structure, researches of Ηe(***) are not systemic. In this paper, we worked out rules and strategy using semantic and grammatical information based on Hierarchical Network of Concepts theory to distinguish the noun, preposition and conjunction of Ηe(***) and identify the beginning and end boundaries of coordinate structures connected by Ηe(***). Compared with Google translator and another rule based translation system we got a relatively satisfied enhancement on the identification precision.