Machine Learning course · Lesson 4
Unsupervised Learning
Discover how data can reveal patterns without predefined target labels. Explore clustering, customer segmentation, anomaly detection, validation and responsible business use through eight interactive concepts.
What Is Unsupervised Learning?
Understand how a model can discover structure in data when target labels are not supplied.
Understand the concept
Unsupervised learning explores data without a known target answer for each record. Algorithms can identify similarities, group observations, or describe unusual patterns. The discovered structure is a starting point for analysis, not automatically a correct business explanation.
Business connection
A retailer can group customers with similar purchasing patterns even when the business has not already assigned customer-segment labels.
Explore customers without preset labels
Select a view to compare how records can be grouped by similarity. The colours are illustrative clusters, not ground-truth categories.
Interactive visual
Colours illustrate a possible grouping based on selected features.
Interactive activity
Apply the concept
A retailer has customer purchase records but no existing segment labels. Which approach best fits the goal of discovering groups?
Frequently Asked Questions
What is unsupervised learning?
Unsupervised learning explores patterns and structure in data without supplied target labels for the task.
What is clustering?
Clustering groups observations according to a chosen definition of similarity.
What is customer segmentation?
Customer segmentation describes groups with shared measured characteristics, which should be interpreted and validated.
Does an anomaly prove fraud?
No. An anomaly is unusual relative to a pattern and should be investigated in context.