MLS-C01 Question 142
Single answerA company wants to develop a machine learning model to predict whether customers will churn (cancel their subscription) based on historical data containing categorical and numerical features. The dataset is imbalanced, with only 10% of the customers labeled as churners. They need a model that can handle the imbalance effectively and provide interpretable predictions for business stakeholders. Which model would be the most appropriate for this use case?
- A
Linear Regression
- B
Logistic Regression with class weighting
- C
K-Means Clustering
- D
Gradient Boosted Trees (e.g., XGBoost) without class weighting
Show answer and explanation
Correct answer: B
Explanation
Logistic Regression with class weighting is the most appropriate choice because the problem involves binary classification with an imbalanced dataset, and the company requires interpretable results. Class weighting ensures the model gives due importance to the minority class (churners), while the interpretability of logistic regression aligns with the business stakeholders' needs.
- A. Incorrect.
Linear Regression is not suitable for binary classification tasks like churn prediction because it is designed for regression tasks (predicting continuous values).
- B. Correct.
Logistic Regression with class weighting is suitable for binary classification tasks, especially when the dataset is imbalanced. The class weighting helps the model focus more on the minority class (churners), and logistic regression provides interpretable coefficients for business stakeholders.
- C. Incorrect.
K-Means Clustering is an unsupervised learning algorithm that groups data into clusters. It is not suitable for a supervised classification task like churn prediction.
- D. Incorrect.
Gradient Boosted Trees (e.g., XGBoost) are powerful models for classification tasks, but without class weighting or other imbalance-handling techniques, they may perform poorly on imbalanced datasets by focusing too much on the majority class.