In one line: Many algorithms — SVMs, k-NN, anything trained with gradient descent, PCA — are sensitive to feature magnitude. A feature ranging 0–1,000,000 will dominate a feature ranging 0–1...
Key Ideas
1Why scale at all?. Many algorithms — SVMs, k-NN, anything trained with gradient descent, PCA — are sensitive to feature magnitude. A feature ranging 0–1,000,000 will dominate a feature r...
Code Examples
from sklearn.preprocessing import StandardScaler, OneHotEncoder
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train) # fit + transform on train
X_test_scaled = scaler.transform(X_test) # transform ONLY on test
encoder ...