Abstraksi teks berbahasa Indonesia menggunakan teori knowledge graph

Abstract

Text summarization system has been developed in order to select several sentences based on certain criteria. However, this method tends to omit some informations contained in unselected sentences. Therefore, a system is needed to get a text abstraction, which represents the whole informations contained in the original text. Knowledge graph offers the way to built such system. Knowledge Graph is a new method in Natural Language Processing (NLP) that represents a knowledge in a form of graph. Using this method, every sentence in the full text version can be represented into graphs which can be further analyzed and gathered into a single graph that describes the meaning of the full text. The result of this research is a procedure to implement knowledge graph theory to get an abstraction of the text.