OPTIMIZING UAM CORPUS FOR TRANSITIVITY PROCESS REGARDING TO COVID-19 NEWS IN INDONESIA FROM FEBRUARY TO MARCH 2020

Abstract

In Systemic Functional Language (SFL), texts are analyzed by their transitivity system; and for the process are well-known as transitivity process. Practically, text parsing takes much time, since the part of text should be parsed manually and analyzed partially. By optimizing UAM corpus tool, the parsing process will get maximum results. UAM concerns on manual as well as semi-automatic annotation because of its lack of accuracy using automatic annotation. In this research, the texts are crawled from The Jakarta Post news article, especially in the form of online data. Those data (texts) are generated by UAM in order to get transitivity properties. The data are taken from February to March which contain Covid-19 issues in Indonesia. The result shows that between the time periods from February to March, the transitivity process regarding to the issue of Covid-19 in Indonesia has changed. It especially refers to the three process; material, mental and verbal process. During that period, material process slightly decreases from 58.72% to 57.59%, which emerges an assumption that people in Indonesia start to decrease their physical activity in March. While in mental process, it significantly elevates from 5.03% to 7.59% which could be assumed that people, including government, tend to think about the issue more seriously. While in the verbal process, it grows by about 2% from 16.11% in February to 18.62% in March which leads to an assumption that people have spread out the issue or the news to the other.