Appendix A: Identifying Anti-vaxxers through Machine Learning Using KNIME Available to Purchase

In this example, we will learn how to use machine learning to identify proponents of the “Anti-Vaxxer” theory through their Twitter behavior. Anti-vaccination, the refusal of parents to vaccinate their infants against common infectious diseases, has been scientifically debunked, but is still propagated by a small but vocal minority of parents in the United States. They claim that vaccination of infants will create autism. The consequence is that in some parts of the United States more than 10% of children are not vaccinated, thereby becoming potential carriers of infectious disease, such as measles, for their peers.¹

We will analyze a dataset of tweets that has been collected in Spring 2015 by a team of students of the COINs seminar at FHNW Brugg and University of Cologne. They gathered all the Tweets containing the words “vaccination,” “vaccinate,” “vaxxer,” “vaccine,” “anti-vaccination,” and “anti-vaxxers.” The resulting Tweets together with information about the tweeters were used to manually classify two sets of Tweets: one belonging to pro-vaxxers and the other belonging to anti-vaxxers.