I. Introduction
Although language resources as datasets are being published, an effective natural language system needs to be preprocessed them. To meet this prerequisite, preparing datasets by linguists is also an alternative approach to improve efficiency in summarizing or translating.