This repository trains the model which is used in the Who said it? web app.
Who said it? is a web app where the user writes a message and finds out whether Bernie Sanders or Donald Trump is more likely to say it. The model is trained by calculating the term frequency–inverse document frequency (tf-idf) vectorization of both politicians tweets. The tweets were obtained using twint. A logistic regression and naive Bayes classifier are then trained and combined in a soft voting ensemble method using scikit-learn.