A FrontPage Online News Analysis: An Opinion Mining Approach

Abstract

The advent of the internet with its attendant democratization of data and deluge of information had given rise to the avalanche of news media agencies. These agencies publish news articles with varying emotional reports especially stories conveying bad sentiments to the public. As major news agencies operate micro-blogging websites and establish their presence on social media channels, the distribution of bad news increases. It has been shown that constant exposure to bad news presented in a body of texts, graphics, and videos/audios contribute to increase in high blood pressure, anxiety attacks, bowel disorders, stroke and/or heart failure. In this work, we presented a sentiment analysis framework to extract news articles from FrontPage of online newspapers and generate contextual wordlists to support positive news broadcasting. Using a set of 12 Nigerian online news channels, we employed a hybrid method of dictionary and corpus-based lexicon approaches to achieve the wordlist derivation. The result advocates for an alternative way of reporting negative news to reduce the adverse impact it has on the masses.