An organization wants to determine why they are losing customers. Which supervised modeling technique would be used to accomplish this task?

Question

Options

• APCA
• BQUEST
• CApriori
• DKohonen

• A
  3% (1)
• B
  8% (3)
• C
  86% (32)
• D
  3% (1)

Explanation

There appears to be an error in the provided correct answer. B (QUEST) is the supervised technique here, not C.

QUEST (Quick, Unbiased, Efficient Statistical Tree) is a decision tree classification algorithm - a supervised technique that uses labeled historical data (customers who left vs. stayed) to build a model predicting and explaining churn. It directly answers "why are customers leaving" by revealing which variables drive the outcome.

Why the others are wrong:

• A. PCA (Principal Component Analysis) - unsupervised dimensionality reduction; it compresses variables but doesn't classify or predict outcomes
• C. Apriori - unsupervised association rule mining (market basket analysis); it finds co-occurrence patterns like "customers who buy X also buy Y," but doesn't use labeled target variables
• D. Kohonen (Self-Organizing Maps) - unsupervised clustering neural network; it groups similar customers but has no labeled outcome to train against

Memory tip: "Supervised = someone already labeled the data." If you have historical records marked churned/not churned, you're doing supervised learning. QUEST is a decision tree variant - decision trees split data on rules that explain outcomes, making them ideal for answering why questions.

I'd recommend double-checking the answer key source, as Apriori is widely taught as an unsupervised method and would be incorrect here.

No community discussion yet for this question.