Feature importance in regression in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding feature importance methods

Which method directly measures the impact of each feature on the prediction error by randomly shuffling its values and observing the change in model performance?

A Feature scaling

B Permutation importance

C Correlation coefficient with target variable

D Coefficient magnitude in linear regression

❓ Predict Output

intermediate

Output of feature importance extraction

What is the output of the following Python code using scikit-learn's RandomForestRegressor to get feature importances?

ML Python

from sklearn.ensemble import RandomForestRegressor
import numpy as np

X = np.array([[1, 2], [3, 4], [5, 6], [7, 8]])
y = np.array([1, 3, 5, 7])
model = RandomForestRegressor(random_state=0)
model.fit(X, y)
importances = model.feature_importances_
print(np.round(importances, 2))

A [0.3 0.7]

B [0.0 1.0]

C [0.5 0.5]

D [0.7 0.3]

❓ Model Choice

advanced

Choosing a model for interpretable feature importance

You want a regression model that provides clear, direct coefficients to interpret feature importance easily. Which model should you choose?

A Support vector regression

B Random forest regression

C Neural network regression

D Linear regression

❓ Metrics

advanced

Evaluating feature importance with model metrics

Which metric change best indicates a feature is important when removed from a regression model?

A Significant increase in mean squared error (MSE)

B Increase in training time

C Decrease in R-squared value

D Decrease in number of features

🔧 Debug

expert

Debugging incorrect feature importance values

You trained a linear regression model but the feature importance (coefficients) are all zero. What is the most likely cause?

A The learning rate was set too high

B The features are perfectly correlated with each other

C The target variable has zero variance (all values are the same)

D The model was trained with too many epochs