Noun Sequences, which frequently appear in Korean text, consist of two or more nouns, and these nouns are related with various semantic relation. The interpretation of noun sequence is to find semantic relation between the nouns in noun sequence. To interpret noun sequence, semantic knowledge about words and relation between words is required.
In this thesis, we propose a method to interpret a semantic relation between nouns in noun sequence. Our work is divided into two parts. In the first part, we extract semantic information from an MRD and corpus using regular expressions. Based on the extracted information, semantic relation of noun sequence is interpreted in the second part. In the second part, we use verb subcategorization information together with the semantic information from an MRD and corpus. In subcategorization, a selectional restriction of a verb is described. So, we can get semantic information about a special category of nouns that can be used as a verb with verbal-suffix "-hada(하다)" from subcategorization information.
Previous researches use semantic knowledge extracted only from an MRD, but our method uses an MRD, corpus, and subcategorization information to interpret noun sequences. Experimental result shows that our method improves the accuracy rate by 40.30% and the coverage rate by 12.73% better than previous researches.