Identifying Supervised Machine Learning Examples
Short answer
A Building a model to predict resolution time using past, labeled support tickets.
Question
Key Takeaways
The exam tests the distinction between supervised and unsupervised learning; the common trap is confusing clustering (unsupervised) with prediction using labels (supervised).
This question tests the fundamental definition of supervised machine learning by requiring candidates to identify scenarios involving labeled training data. It establishes that predicting outcomes from historical, labeled inputs is the core characteristic of supervised learning.
Common mistake: Candidates often select Option B or C because they involve analyzing data, but fail to notice the absence of 'labels' or 'known outcomes' in those options, which are required for supervised learning.
Expert Analysis
Why the Answer Is Correct
Supervised machine learning relies on a labeled dataset where the algorithm learns to map input variables to output variables based on example input-output pairs. Option A describes building a model to predict resolution time using past, labeled support tickets. Since the tickets are explicitly described as "labeled," this represents a classic supervised learning task (regression or classification) where the model learns from known outcomes to predict future ones.
Why the Other Options Are Wrong
Options B, C, and D describe unsupervised learning techniques. Option B (identifying unknown customer segments) and Option C (finding user behavior clusters) are examples of clustering, an unsupervised method used to find hidden patterns in unlabeled data. Option D (automatically identifying recurring topics) typically involves Natural Language Processing techniques like topic modeling, which is also generally unsupervised unless specific topic labels are provided for training.
Community Comment Notes
Community feedback supports Option A, with users noting that supervised models premise their operation on using pre-labeled data. One commenter explicitly stated, "Supervised models have the premisse of using pre labeled data," reinforcing that the presence of labels is the deciding factor.
Exam Strategy
When evaluating machine learning questions, always look for keywords like 'labeled,' 'historical outcomes,' or 'prediction based on known data' for supervised learning. If the scenario mentions 'clustering,' 'segmentation,' or 'discovering patterns' without reference to known labels, it is likely unsupervised learning.
Frequently Asked Questions
Why is identifying customer segments not supervised learning?
Can topic modeling be supervised?
Discussion 1 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 1
-
gsilva 👍 2 Answered ASupervised models have the premisse of using pre labeled data.
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.