Sentiment Analysis of Cak Nun on Youtube and Online News: Multinomial Naive Bayes for Positive, Neutral, and Negative Perspectives

Abstract

In the era of digitalization, which is saturated with abundant information, sentiment analysis has emerged as a crucial tool for comprehending and addressing the intricate nature of public opinion. The copious quantity of textual data generated by media platforms such as YouTube and online news outlets offers valuable insights into the viewpoints held by the public on a diverse range of topics and public figures. Consequently, sentiment analysis plays a pivotal role in discerning the prevailing direction of sentiment, be it positive, negative, or neutral. This article focuses on the sentiment analysis of Cak Nun, a highly esteemed cultural figure and poet from Indonesia. The utilized data consists of titles from YouTube and pertinent articles from online news sources that pertain to Cak Nun. The chosen methodology is the Multinomial Naive Bayes with CountVectorizer feature selection. By employing the Multinomial Naive Bayes, the patterns present within the text are learned to classify the textual data. At the same time, the CountVectorizer identifies the critical aspects within the evaluations of YouTube titles and online news articles. The resulting accuracy achieved is 82.11%, thereby indicating the effectiveness of the Multinomial Naive Bayes in accurately classifying sentiment. Overall, the model produces favorable outcomes, although there remains room for improvement in its ability to handle negative sentiment, as the precision, recall, and F1-Score values for negative sentiment are slightly lower than those for other sentiments. This sentiment analysis is anticipated to yield considerable advantages for the public figure known as 'Cak Nun,' as well as for the wider public and researchers in terms of further advancements and progress in the future.