MLS-C01 · Question #200
An ecommerce company sends a weekly email newsletter to all of its customers. Management has hired a team of writers to create additional targeted content. A data scientist needs to identify five…
The correct answer is D. The XGBoost model is a supervised machine learning algorithm. Train a k-means model. The XGBoost model fails here for two reasons: (1) it is a supervised learning algorithm requiring labeled training data (known categories), but customer segmentation has unknown/unlabeled groupings; and (2) even if retrained for multi-class output, it cannot discover natural…
Question
An ecommerce company sends a weekly email newsletter to all of its customers. Management has hired a team of writers to create additional targeted content. A data scientist needs to identify five customer segments based on age, income, and location. The customers’ current segmentation is unknown. The data scientist previously built an XGBoost model to predict the likelihood of a customer responding to an email based on age, income, and location. Why does the XGBoost model NOT meet the current requirements, and how can this be fixed?
Options
- AThe XGBoost model provides a true/false binary output. Apply principal component
- BThe XGBoost model provides a true/false binary output. Increase the number of classes
- CThe XGBoost model is a supervised machine learning algorithm. Train a k-Nearest-
- DThe XGBoost model is a supervised machine learning algorithm. Train a k-means model
How the community answered
(24 responses)- A4% (1)
- D96% (23)
Explanation
The XGBoost model fails here for two reasons: (1) it is a supervised learning algorithm requiring labeled training data (known categories), but customer segmentation has unknown/unlabeled groupings; and (2) even if retrained for multi-class output, it cannot discover natural groupings without labels. The correct fix is k-means, an unsupervised clustering algorithm that can discover 5 natural customer segments from age, income, and location without labeled data. k-Nearest Neighbors (option C) is also supervised learning, so it has the same fundamental problem. Principal Component Analysis (PCA) in option A is a dimensionality reduction technique, not a clustering solution.
Topics
Community Discussion
No community discussion yet for this question.