New machine learning model for the early detection of fake news

What is it about?

Fake news has the potential to incite tensions and disrupt the peace. Being able to detect fake news, has thus become very important. Most methods to detect fake news focus on how it spreads via social media. But social media spread is the final stage of the news cycle. This study proposes a machine learning model to detect fake news early in the news cycle. The method studies the content of news articles at four language levels. The first level looks at how often certain words are used in the article. The second focuses on the arrangement of certain words and phrases. The third examines the meanings of words and the sentiments they express. The fourth level evaluates if the article contains particularly persuasive words or phrases. The authors then tested the method on articles from two news websites. They found that their method could outperform some of the latest fake news detection methods.

Featured Image

Photo by Jorge Franganillo on Unsplash

Photo by Jorge Franganillo on Unsplash

Why is it important?

By the time fake news reaches social media, it has already been consumed by many people. Studies have also shown that it is difficult to change one's thoughts after fake news has already gained their trust. This makes it important to catch fake news at an early stage of the news cycle. This study proposes an accurate method to do just that. It also explores potential patterns in fake news. The model further makes it easier for fake news feature engineers to understand how the machine learning algorithm detects fake news. The paper also studies the relationships between fake news, deception and disinformation, and clickbait. Thus, future work on fake news detection can build upon this research. KEY TAKEAWAY: Detecting fake news before it can spread on social media is an important way to curb its ill effects. The model proposed in this study can detect fake news based on its writing style, quality, and article structure.