Classification of Public Figures Sentiment on Twitter using Big Data Technology

Abstract

Public figures often receive widespread public attention because they can exert a meaningful influence. On Twitter, the users can freely express their opinion through tweets. There are about 456,000 tweets sent in a minute which with this large and diverse number will make Big Data. Big Data has valuable potential for better decision-making. This large amount of tweet data can yield valuable information through sentiment analysis. This study aims to conduct a sentiment classification of Indonesian public figures using Twitter's data. This study used 1,034,329 tweets collected from Twitter in the period November 2021 until March 2022. Tweet classification is done by building a classification model using the Bidirectional Long Short-Term Memory algorithm. Sentiment toward public figures in Indonesia is 45.98% negative sentiment, 28.04% positive sentiment, and 25.98% neutral sentiment resulting from this study. The highest positive sentiment is obtained by public figures when there is content or news that is relevant to the public figure, while the highest negative sentiment is obtained when there is content or news that contradicts the image of the public figure.