Penambahan modul pembentukan kamus word graph kata benda pada sistem Aplikasi bogordelftconstruct

Abstract

Knowledge Graph is a new method in Natural Language Processing that is used to describe human language and displaying it into a graph form. BogorDelftConStruct is a tool to analyze Indonesian text, developed by Deni Romadoni (2009). The tool has limited feature, opening many opportunities to add some other features. The goal of this research is to develop a noun word graph dictionary module and add it in BogorDelftConStruct. The patterns of the noun word-graph are based the research of Saleh (2009). In this research, of the 20 patterns in Saleh (2009) only 13 patterns are used, since some of word graph patterns which have the same affix and also have the different meaning. As many as 144 nouns were used in scenario testing, with 1 error was generated (99,33% accuracy).