Training Logistic Regression, Naïve Bayes and Support Vector Machine classifiers on a user’s historical tweets leads to accurate classification of accounts most at risk of fixation with the conspiracy.