MLS-C01 Question 152
Single answerA data scientist is tasked with building a machine learning model to predict customer churn for a subscription-based service. The stakeholders of the project insist that the model's predictions must be explainable and interpretable for business decision-making. Which type of model should the data scientist prioritize, and why?
- A
Decision Trees, because they provide clear and interpretable decision boundaries.
- B
Deep Neural Networks, because they automatically learn complex patterns in the data.
- C
Support Vector Machines, because they maximize the margin between data points.
- D
Ensemble models like Random Forests, because they combine multiple models for improved accuracy and interpretability.
Show answer and explanation
Correct answer: A
Explanation
For applications requiring explainability, models like Decision Trees are favored because they provide transparent decision-making processes that can be easily understood by stakeholders. While other models may offer better performance in some cases, their lack of interpretability makes them less suitable for scenarios where understanding the intuition behind the model is critical.
- A. Correct.
Decision Trees are inherently interpretable and provide a clear path to understanding the decisions made by the model, which aligns with the stakeholders' requirement for explainability.
- B. Incorrect.
Deep Neural Networks are powerful but often referred to as 'black-box' models due to their lack of interpretability, which does not meet the stakeholders' needs for explainability.
- C. Incorrect.
Support Vector Machines, while effective in certain contexts, do not inherently provide interpretable decision-making processes, making them less suitable for this use case.
- D. Incorrect.
Ensemble models like Random Forests improve accuracy but sacrifice interpretability because they aggregate the outputs of many individual decision trees, making the overall decision process less transparent.