Machine Learning

Common Models

What to try first on real-world tabular problems.

Quick Guidance

  • Linear / Logistic Regression: fast + interpretable baseline.
  • Random Forest: strong default for tabular; less tuning.
  • Gradient Boosting: often top performance on tabular; more tuning.
  • SVM / KNN: can work well on smaller datasets; need scaling.

Starter Example

from sklearn.ensemble import RandomForestClassifier
model = RandomForestClassifier(
  n_estimators=400,
  random_state=42,
  n_jobs=-1
)
model.fit(X_train, y_train)