Understanding the Role of TF-IDF in Scikit-learn's Text Classification Pipeline and Overcoming Accuracy Issues with Smoothing Techniques