Perbandingan Metode Knowledge Graph dan Semantic Network untuk Analisis Teks Berbahasa Indonesia.

Abstract

Knowledge is something that related to the learning process. Knowledge is obtained from various sources, for example from text. In many cases the text contains ambiguity. Knowledge graph and semantic network are mathematical concepts that can be used to resolve the ambiguity. The objective of this research was to compare the two methods to analyze an Indonesian text. The comparison was done by emphasizing on the aspect of process and on the aspect of graph structure. Research done by making graph on sample sentences in Indonesian text using both methods. Based on the aspect of the process, steps of the knowledge graph is more complex, but its automatic procedure is easier than the semantic network. It is easier because the knowledge graph has limited number of concepts, whereas the semantic network has an unlimited number of concepts. Based on the aspect of graph structure, relation labels of the knowledge graph has no meaning, it refers to the eight binary relationships, whereas the relation labels of the semantic network have meaning to describe relation of the concepts. Both methods have advantages and disadvantages but from this research it has been known that the knowledge graph is more feasible for an automatic procedure.