6 Clustering Free Online FlashCards

What is clustering?

Clustering groups examples by similarity in their features without using target labels.

How does unsupervised learning differ from supervised learning?

Supervised learning uses examples with target labels or values; unsupervised learning looks for structure in examples without them.

What is the main goal of clustering?

It aims to make examples within each cluster more alike than examples in different clusters, according to its assumptions and similarity measure.

How can clustering help explore data?

Clustering can reveal possible patterns or subgroups to investigate before deciding what to explore further.

Which data choices can change clustering results?

Results can change with the selected features, their representation or scaling, and the definition of similarity.

Do all clustering methods require a preset number of clusters?

Some methods require the number of clusters to be chosen in advance; others find groups using different criteria.

What does a cluster ID tell you about a group?

A cluster ID is just a label assigned by the algorithm; it does not make one cluster inherently more important or meaningful.

What should you check when judging whether clusters are useful?

Check whether groups are understandable, useful for their purpose, and reasonably consistent when appropriate choices or samples change.

What is a cluster?

A clustering method assigns examples to groups called clusters.

How can clustering summarize data?

It can summarize many examples using a smaller number of group profiles.

How should you interpret a cluster?

Examine members’ original features and describe the patterns that distinguish each group.

Do clusters prove that fixed, natural categories exist?

No. A cluster may reflect an unhelpful feature, a data-preparation choice, or a pattern specific to the dataset.