Sentiment analysis of media behavior during covid-19 pandemic

Abstract

Online media applications are the most popular and convenient method to get the news across the globe. There were some instances of unethical behavior by the media when reporting misleading and fake news. This research presents a method to perform a sentiment analysis specifically designed to work with COVID-19 social media news, taking into account their structure, length, and specific keywords related to the language used. Initially, the data were revealed through pre-processing of media news posts to normalize the language, by cleaning text data and generalizing the text to express the sentiment. The text data collection from COVID-19 social media posts and news websites was investigated in the data processing stage. This paper proposes the use of a text blob library as an approach for spot modification in the polarity of the sentiment expressed. Adopting a sentiment analysis through the social audience about the COVID-19 pandemic, their emotions, and moods are split by words to reveal how people feel exactly about the impact of COVID-19 media news. This research contributes to building a training model to identify the sentiment and it will enhance the sentiment classification performance, irrespective of the domain and distribution of the test set.