Why look at feature importance?

It tells you which features drive the prediction, so you can explain the model and drop features that add little. A clearer model and fewer useless features are the payoff.

How do I check a feature-importance ranking?

Build data where you already know which feature should matter, then confirm the ranking puts it where you expect.

What libraries does this problem use?

Python, numpy, and scikit-learn's random forest classifier. You return every feature name exactly once, ordered most important first.

pick out the important features

Your task

Knowing "which feature really matters for the prediction" is central to understanding a model and pruning useless features. A random forest is an ensemble of decision trees, and it also reports a per-feature importance. Complete rank_features(X_train, y_train, X_test, y_test, feature_names). X_train, y_train training features and labels. X_test, y_test a held-out set of features and labels. feature_names the name of each column (feature). Train a random forest and return the feature names as a list in order of importance (most important first). Example ['income', 'age', 'zipcode'] means income is the most important.

Ground rules

•X_train and X_test are 2-D float arrays; feature_names is a name list as long as the column count.
•the model is a random forest classifier.

Not required

•automated feature selection or hyperparameter search.
•comparing against other models (boosting, etc.).

▶About this challenge

This challenge asks you to complete a function that trains scikit-learn's random forest and returns the feature names as a list ordered by importance, most important first. It takes training and held-out feature sets with their labels plus the column names, and the returned list must contain every name exactly once. It practices reading which features matter most to a model's prediction.

Knowing which feature really matters for a prediction is central to understanding a model and pruning features that don't pull their weight. A random forest is an ensemble of decision trees, and it can report a per-feature importance. This problem produces a feature ranking with scikit-learn and verifies it against data you control rather than trusting a plausible-looking order.

Why look at feature importance?: It tells you which features drive the prediction, so you can explain the model and drop features that add little. A clearer model and fewer useless features are the payoff.
How do I check a feature-importance ranking?: Build data where you already know which feature should matter, then confirm the ranking puts it where you expect.
What libraries does this problem use?: Python, numpy, and scikit-learn's random forest classifier. You return every feature name exactly once, ordered most important first.