{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

Clustering - Clustering KwokLeung Tsui Industrial Systems...

Info iconThis preview shows pages 1–22. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Clustering Kwok Leung Tsui Industrial & Systems Engineering Georgia Institute of Technology
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
2 Statistical Learning Methods Supervised Learning • Statistical data mining techniques creating a functions from a training dataset. • Prediction or classification for unknown datasets. Unsupervised Learning Statistical data mining techniques grouping or partitioning datasets (clustering). • Visualization or variable reduction. • Association rules or correlation analysis.
Background image of page 2
3 Clustering • “Clustering is the most frequent data mining task.” KDnuggets News, 6/11/01 • Provides organizing and simplifying structure for large, complex data sets • Global vs. local analysis – Zero in on clusters for local analysis – Combine results for global perspective
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
4 Learn a method for predicting the instance class from pre labeled (classified) instances Classification Example Classification Models
Background image of page 4
5 Find “natural” grouping of data given un labeled data Clustering Example
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
6 Continuum of Problems
Background image of page 6
7 Basic Types of Algorithms • Hierarchical – Tree of objects; dendrogram – Agglomerative or divisive approaches • Partitioning – Distinct, non-overlapping groups – Iterative shuffling of objects •O t h e r – Fuzzy – Overlapping
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
8 Hierarchical Methods • Start with each object alone in its own group • Compute all pairwise “distances” among groups • Combine closest pair of groups • Recompute intergroup distances – Maximum, minimum, or average distance rules, e.g. • Continue until all objects united into one group • Analyze the resulting hierarchical tree – major branch of tree could be a cluster of interest
Background image of page 8
9
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
10
Background image of page 10
11 Clustering Methods • Hierarchical Clustering 1) Neighbor joining clustering (Single linkage clustering) 2) Complete linkage clustering 3) The average linkage clustering 4) Within-group clustering 5) Ward’s method • Non-Hierarchical Clustering 1) k-means clustering 2) Self-organizing maps
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
12 A B C D Dist ABCD A2 0 7 2 B1 0 2 5 C3 D Distance Matrix Initial Data Items Hierarchical Clustering Methods
Background image of page 12
13 A B C D Dist ABCD A2 0 7 2 B1 0 2 5 C3 D Distance Matrix Initial Data Items Hierarchical Clustering Methods
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
14 Current Clusters Single Linkage Dist ABCD A2 0 7 2 B1 0 2 5 C3 D Distance Matrix A B C D 2 Hierarchical Clustering Methods
Background image of page 14
15 Dist AD B C AD 20 3 B1 0 C Distance Matrix Current Clusters Single Linkage A B C D Hierarchical Clustering Methods
Background image of page 15

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
16 A B C D Dist AD B C AD 20 3 B1 0 C Distance Matrix Current Clusters Single Linkage Hierarchical Clustering Methods
Background image of page 16
17 Dist AD B C AD 20 3 B1 0 C Distance Matrix Current Clusters Single Linkage A B C D 3 Hierarchical Clustering Methods
Background image of page 17

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
18 Dist AD C B AD C 10 B Distance Matrix Current Clusters Single Linkage A B C D Hierarchical Clustering Methods
Background image of page 18
19 A B C D Dist AD C B AD C 10 B Distance Matrix Current Clusters Single Linkage Hierarchical Clustering Methods
Background image of page 19

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
20 Dist AD C B AD C 10 B Distance Matrix Current Clusters Single Linkage A B C D 10 Hierarchical Clustering Methods
Background image of page 20
21 A B C D Dist AD CB AD CB Distance Matrix Final Result Single Linkage Hierarchical Clustering Methods
Background image of page 21

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Image of page 22
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

Page1 / 70

Clustering - Clustering KwokLeung Tsui Industrial Systems...

This preview shows document pages 1 - 22. Sign up to view the full document.

View Full Document Right Arrow Icon bookmark
Ask a homework question - tutors are online