Peer Reviewed Journal via three different mandatory reviewing processes, since 2006, and, from September 2020, a fourth mandatory peer-editing has been added.
In this paper, we present a named entity recognition model for Korean Language. Named entity recognition is an essential and important process of Question Answering and Information Extraction system. This paper proposes a HMM based named entity recognition using compound word construction principles. In Korean, above 60% of NE (Named-Entity) is a compound word. This compound word may be consisted of proper noun, common noun, or bound noun, etc. There is an intercontextual relationship among nouns which consists NE. NE and surrounding words of NE have a contextual relationship. For considering these relationships, we classified nouns into 4 word classes (Independent Entity, Constituent Entity, Adjacent Entity, Not an Entity). With this classification, our system gets contextual and lexical information by stochastic based machine leaning method from a NE labeled training data. Experimental result shows that this approach is better approach than rulebased in the Korean named-entity recognition.
