Add tf-idf vectorization steps and then compare different ML models. Evaluate the best ML model and evaluate using a confusion matrix