MLS-C01 Question 128
Single answerA company operates an e-commerce platform and wants to reduce customer churn by identifying users who are likely to stop using the platform within the next 30 days. They want to take proactive actions, such as offering discounts or personalized recommendations, to retain these users. How should you frame this business problem as a machine learning (ML) problem?
- A
Frame the problem as a classification task to predict whether a user is likely to churn or not within the next 30 days.
- B
Frame the problem as a regression task to predict the exact number of days until a user churns.
- C
Frame the problem as a clustering task to group similar users based on their activities and identify churn patterns.
- D
Frame the problem as a reinforcement learning task to optimize the rewards for retaining users.
Show answer and explanation
Correct answer: A
Explanation
Framing a business problem as an ML problem involves identifying the most appropriate ML task that aligns with the business objective. In this case, the goal is to identify users likely to churn within a specific time frame, making it a binary classification problem. This approach provides actionable insights, enabling the company to take targeted interventions and reduce churn.
- A. Correct.
Correct: Predicting whether a user will churn within the next 30 days is a binary classification task. It directly addresses the business objective of identifying at-risk users, allowing the company to take targeted actions.
- B. Incorrect.
Incorrect: Regression tasks focus on predicting continuous numerical variables. While predicting the exact number of days until churn might provide some insights, it does not directly align with the business goal of identifying users likely to churn within a specific time window.
- C. Incorrect.
Incorrect: Clustering groups similar users but does not directly predict churn. It could be used as a preliminary step to understand user behavior, but it does not solve the core problem of identifying specific users likely to churn within 30 days.
- D. Incorrect.
Incorrect: Reinforcement learning optimizes sequential decision-making processes, such as dynamic pricing or recommendation systems. However, it is not the appropriate approach for predicting churn, which is a supervised learning problem.