Modelling Veracity of Football Player Trade Rumours on Twitter Using Naive Bayes Algorithm

Abstract

Twitter today has become one of the most influential social media application in our world. Twitter is a source of a plethora of data contributed by its millions of users. Twitter is a popular choice for journalists reporting about football to disseminate information about impending player transfers. Football has become very popular amongst people and draws a lot of social media engagement towards news of player trading. This has unfortunately given rise to several "in the know" social media accounts that propagate fake news to exploit fundamental flaws in social media ranking applications. This paper attempts to gather data about specific words most commonly used during the period of a player transfer occurring and model it using the Naive Bayes algorithm to determine whether a player transfer has occurred given the choice of words expressed in a tweet whilst comparing its results to models in use for detecting the veracity of transfer rumours.