Accuracy Analysis of Pasang Aksara Bot using Finite State Automata Transliteration Method

摘要

Bali is a world tourism destination for its cultural riches. Balinese culture is a legacy of the world’s riches that must be preserved. In the development of the education world and technology today, there have been many efforts done to preserve Balinese culture such as researches used to document Balinese cultures. One study for the preservation of Balinese culture in the field of technology is the Balinese script transliteration application. Balinese script is one of the traditional Indonesian archipelago that created in Bali. Because its use is limited to a narrow scope in everyday use, most Balinese script has been replaced with Latin letters. Therefore this research is focused on preserving Balinese script using robot technology. The robot that developed in this research is a robot that can write Balinese script using a design resembles a human arm. Its connected to the computer using USB and mDraw a graphics program developed based on open source python. mDraw will give input to the robot in .svg format that generated by the Pasang Aksara transliteration application. This .svg or vector as input mDraw will give command to the robot to write the transliteration result on a piece of paper. In this paper a new idea has been proposed to implement the robot on cultural preservation and Balinese script education. This robotic arm was successfully tested and the result are 60% (21 glyph) categorized as very good, 37.14% (13 glyph) categorized as good, and 2.86% (1 glyph) enough. And for the Pasang Aksara transliteration aplication obtained the result of accuracy 75,73 % (of 861 words) for the correct test count and 24,7 % (of 276 words) for wrong test, count from 1137 word on thirteen kinds of special word in transliteration rules. For future development, replacement servo motors are located on the pen to make the process on the ups and downs of the pen smoother for more accurate Balinese script writing by the robot. The FSA method used is still constrained by some special words that failed to be transliterated correctly, so it is necessary to develop a certain algorithm so the accuracy of transliteration can reach 100%.