🕊️ Naive Bayes

A fast, probability-based algorithm that powered early spam filters.

🙈 Why is it "Naive"?

It assumes every feature (word) is completely independent of every other feature. It thinks the word "New" and the word "York" have no connection! Despite this, it is shockingly accurate for text.

🐍 Python Implementation

from sklearn.naive_bayes import MultinomialNB
from sklearn.feature_extraction.text import CountVectorizer

# Simple text data
emails = ["Get free money now", "Hi mom how are you", "Free viagra click here"]
labels = [1, 0, 1] # 1 = Spam, 0 = Normal

# Convert text to word counts
vectorizer = CountVectorizer()
X = vectorizer.fit_transform(emails)

# Train Naive Bayes
model = MultinomialNB()
model.fit(X, labels)

# Test a new email
test = vectorizer.transform(["free stuff here"])
print("Is it spam?", bool(model.predict(test)[0]))