📖 How Machine Learning Works
Not magic. Maths + patterns + data.
1️⃣
Step 1: Collect Data — Get 100 emails. Label each: SPAM or NOT SPAM. More data = smarter AI.
2️⃣
Step 2: Extract Features — Turn text into NUMBERS. Count exclamation marks, "FREE", links, ALL CAPS. Each = a feature.
3️⃣
Step 3: Find Patterns — Algorithm finds rules: "3+ exclamations AND 'FREE' = spam 95% of the time." Decision tree.
4️⃣
Step 4: Test It — Give it UNSEEN emails. Correct answers = accuracy. 90% = 9 out of 10 right.
5️⃣
Step 5: Improve — Add data, add features, try algorithms. 85% → 92% → 97%. That's learning.
🐍 The Actual Code (10 lines!)
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import train_test_split
# Load data
X = [[3,1,1,0], [0,0,0,1], ...]
y = ["spam", "ok", ...]
# Train
model = DecisionTreeClassifier()
model.fit(X_train, y_train)
# Predict
model.predict([[5,1,1,0]]) # → "spam"
💡
10 lines. That's all. The student imports scikit-learn, trains a model, and predicts. The library does the maths — the student understands the LOGIC.