This thesis discusses the generation of Korean morpheme in PIVOT machine translation systems. Given the syntactic and semantic structure, the meaning that is represented in these structures is converted to Korean surface sentence.
At first, semantic features are ordered for the predicates(verb ,adjective). Then are sequentially converted to surface words. Syntactic features are converted to affixes for the nouns, and the postpositions (josa) are selected according to grammar cases of the nouns.
In this thesis, the generation process is independent with the analysis process, and Korean surface representation is generated by investigating the meaning represented in semantic and syntactic structures. So, this thesis is applied to interface of Korean natural language processing system as well as machine translation systems.