LANGUAGE ATTITUDE AND SENTIMENT ANALYSES IN GETTING THE INSIGHTS TOWARDS COVID-19’S OMICRON VARIANT NEWS

Abstract

Omicron variant has been massively reported on Indonesian mass media following the spread of other previous variants during Covid-19 pandemic. This research combines computer science and linguistics to analyze the news on the variant. It implemented quantitative research using computational algorithm by collecting the titles of the news from Indonesian mainstream online mass media. Sentiment Analysis (SA) was applied to obtain the sentiments, opinion, and subjectivities of the texts along with topic modeling in classifying the topics. The words in the headline news titles were used as the data and grabbed by Python programming language. A criterion-based sampling was employed in to select the relevant data and to formulate the criteria in the research methodology. The results were filtered to ‘Omicron’ keyword for SA processing by the Azure Text Sentiment Analysis tool. The results of SA, as computational research, was then confirmed with Attitude Analysis (AA) from the perspective of Systemic Functional Linguistics. AA classified the words into affect, judgment, and appreciation as the attitude construed in English text. This research provides SA as the insights of Omicron issue. The presence of AA extracts the words into bipolar senses of human’s meaning interpretation. AA is important to straighten SA findings. SA has contextual meaning problem and requires study on its words classified in ‘neutral’ which are then confidently directed into positive or negative meanings by AA. It is found that there are different dynamics by SA and AA findings as they reflect particular meanings. Besides their difference, SA is useful for managing overload data into fast policy making whereas AA makes sure the acceptable meanings to people. In this case AA corrects the bias occurring from SA.