ABSTRACT
As the number of individuals using Facebook, Twitter, and other social media platforms has grown, news travels at a fast pace across millions of individuals worldwide. These platforms do not verify the users or their posts. People fall prey to various kinds of fake news. Therefore, there exists a need to develop a system that allows users to determine whether a content is real or fake. Machine learning (ML) classifiers have been used by researchers; however, most existing algorithms are supervised, requiring a significant amount of time and effort. Also, they mainly focus on controversial topics and tend to expose a distinctive type of news favoring certain news agencies. The impact of spreading fake news is widespread, affecting everything from formation of prejudiced opinions to changing election results in favor of certain leaders. In this chapter, we conduct a comprehensive algorithmic comparison of six machine learning algorithms and contrast it with other models. From results, it is observed that decision tree algorithm with 99% accuracy outperforms all other ML algorithms in intra-model comparison. In inter-model comparison, the algorithm outperforms validated existing state-of-the-art models by 10 points over the next best model.
