Finding useful signal in text.
I built this project to understand how text becomes a signal a model can learn from. The pipeline handles tokenization, stop-word removal, and normalization, then converts text to numerical features with TF-IDF.
I compared Logistic Regression and Support Vector Machine classifiers using accuracy, precision, recall, and F1. I also explored frequent words in a Purdue Global website corpus and used NLTK's WordNet to investigate synonym and antonym relationships.
The useful part was working through the whole pipeline: preparing the text, choosing a representation, and checking how the models behaved.

