The pipeline is simple but effective: I clean the text, convert it into TF-IDF features, and apply classical machine-learning models. I experimented with logistic regression, gradient boosting, and decision trees. The best performing model was a decision tree with about 99% accuracy on the dataset I used. I wrapped the model inside a Streamlit application so users can paste text and get instant predictions. This project helped me understand text preprocessing, vectorization, and model evaluation and how to build a usable ML application end-to-end.”