📧 Build a Spam Email Detector with Naive Bayes

Machine learning models need numbers, not words. TF-IDF converts each message into a vector of word importance scores.

from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer

X = df['text']
y = df['label']

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

vectorizer = TfidfVectorizer(stop_words='english', max_features=5000)
X_train_vec = vectorizer.fit_transform(X_train)
X_test_vec  = vectorizer.transform(X_test)

print(f'Training samples: {X_train_vec.shape[0]}')
print(f'Features (words): {X_train_vec.shape[1]}')

Training samples: 4459
Features (words): 5000

MultinomialNB is the classic choice for text classification — fast, accurate, and interpretable.

from sklearn.naive_bayes import MultinomialNB
from sklearn.metrics import accuracy_score, classification_report

model = MultinomialNB()
model.fit(X_train_vec, y_train)

y_pred = model.predict(X_test_vec)

print(f'Accuracy: {accuracy_score(y_test, y_pred):.4f}')
print()
print(classification_report(y_test, y_pred))

Accuracy: 0.9739

              precision    recall  f1-score   support
         ham       0.97      1.00      0.98       965
        spam       1.00      0.79      0.88       150

    accuracy                           0.97      1115

def predict_spam(message):
    vec = vectorizer.transform([message])
    prediction = model.predict(vec)[0]
    probability = model.predict_proba(vec)[0]
    spam_prob = probability[model.classes_.tolist().index('spam')]
    label = '🚫 SPAM' if prediction == 'spam' else '✅ HAM (not spam)'
    print(f'{label}  (spam probability: {spam_prob:.1%})\n  "{message[:60]}..."')

predict_spam("Congratulations! You've won a FREE iPhone. Click here NOW to claim!")
predict_spam("Hey, are we still on for lunch tomorrow at 1pm?")
predict_spam("URGENT: Your bank account has been suspended. Verify now!")
predict_spam("Can you send me the meeting notes from this morning?")

🚫 SPAM  (spam probability: 99.8%)
  "Congratulations! You've won a FREE iPhone. Click here NOW to..."

✅ HAM (not spam)  (spam probability: 0.1%)
  "Hey, are we still on for lunch tomorrow at 1pm?..."

🚫 SPAM  (spam probability: 97.4%)
  "URGENT: Your bank account has been suspended. Verify now!..."

✅ HAM (not spam)  (spam probability: 0.3%)
  "Can you send me the meeting notes from this morning?..."