Skip to main content

⚖️ Model Selection

How do you choose which algorithm to use?

🛠️ The Cheat Sheet​

  • Tabular Data? Use XGBoost, LightGBM, or CatBoost.
  • Images/Text? Use Deep Learning (PyTorch).
  • Need hyper-fast predictions? Use Naive Bayes or Logistic Regression.
  • Need to explain why? Use a simple Decision Tree.

🐍 Python Implementation (GridSearch)​

We use GridSearchCV to automatically test hundreds of models to find the best one!

from sklearn.model_selection import GridSearchCV
from sklearn.ensemble import RandomForestClassifier

# Define the models and settings we want to test
param_grid = {
    'n_estimators': [10, 50, 100],
    'max_depth': [None, 3, 5]
}

# The automated tester
grid = GridSearchCV(RandomForestClassifier(), param_grid, cv=3)
# grid.fit(X_train, y_train)

# print("Best settings found:", grid.best_params_)