Unsupervised Learning

Unsupervised learning is a family of machine learning methods that work on data with no labels. The algorithm looks for structure on its own, such as clusters, patterns, or low-dimensional summaries.

Why it matters

Labeled data is expensive, while unlabeled data is abundant. Unsupervised methods support customer segmentation, anomaly detection, topic discovery, and data compression, and are often used to explore a dataset before building a supervised model.

Example

K-means clustering groups customers by behavior:

Input:  monthly spend, visits per month
Output: cluster 1 = frequent low spenders
        cluster 2 = occasional big spenders
        cluster 3 = dormant

You then interpret and name the clusters yourself.

Common mistakes

• Treating clusters as real categories when they may be artifacts of the chosen settings.
• Not scaling features, so one large-valued column dominates.
• Expecting a single objectively correct result; there is usually no ground truth to score against.

Related

• Supervised learning
• Dataset
• Data science
• Machine learning

This page is general information written by our editorial team and updated when documentation changes. Servers and providers differ, so test changes safely first. See our editorial policy and disclaimer.