Implementasi Knowledge Graph untuk Abstraksi Teks

Abstract

Knowledge is an important factor in human life. Knowledge can be obtained from various sources and one of them is from a text. The text used in this study is Indonesian text on food security. To gain knowledge in the text, we use a so called Knowledge Graphs (KG) method. This study implements KG method in selecting sentences to represent the knowledge contained in the text. The selection of the sentences in the text is based on the number of nouns in each sentence and position of main sentence in the paragraph. Generally, we select about half of the number of the sentences in each paragraph. In this research, we used four texts on food security; three texts for experimenting and one text for testing. Since some of the sentences were discarded from the paragraph, the selection process caused loss of information in the text. However, the method is relatively simple since only few sentences need to be transformed into a graph. For the long time goal, KG is believed to be a potential method for developing an abstraction system for Indonesian text.