Update README.md

README.md

@@ -1,3 +1,32 @@
----
-license: apache-2.0
----
+# Spam Detection — English (Naive Bayes)
+
+A lightweight spam/ham text classifier for English messages, built with a 
+custom preprocessing pipeline (tokenization, stopword removal, lemmatization) 
+and TF-IDF features feeding into a Multinomial Naive Bayes classifier.
+
+## Model Details
+- **Architecture:** TF-IDF + Multinomial Naive Bayes (scikit-learn Pipeline)
+- **Preprocessing:** Custom transformer — hashtag/punctuation removal, 
+  tokenization (NLTK), stopword removal, lemmatization (WordNet)
+- **Hyperparameters:** Tuned via GridSearchCV (alpha smoothing)
+- **Accuracy:** 99.4% on held-out test set
+
+## Intended Use
+Binary spam classification for English text messages/emails. Part of a 
+multilingual spam detection system that automatically routes text to a 
+language-specific model (English or Arabic) based on detected language.
+
+## How to Use
+\```python
+import joblib
+
+model = joblib.load("spam_eng_nb.joblib")
+prediction = model.predict(["Congratulations! You've won a free prize, click here now"])
+print(prediction)  # 1 = spam, 0 = ham
+\```
+
+## Limitations
+- Trained on a specific dataset distribution; may not generalize well to 
+  domains very different from training data (e.g. highly technical or 
+  slang-heavy text)
+- Naive Bayes assumes word independence — does not capture context or word order