hrwhisper/SpamMessage

Implements multiple classification algorithms (Perceptron, Logistic Regression, Naive Bayes, SVM) with both custom implementations and scikit-learn wrappers, using jieba for Chinese tokenization and bag-of-words feature representation. The pipeline includes separate training (cross-validation in test.py) and inference phases, with trained models serialized for reuse. Accepts unlabeled SMS files via command-line interface and outputs binary spam/non-spam predictions.

201 stars. No commits in the last 6 months.