Cluster Analysis part2

1. Cluster Analysispart2 Dr. Bernard Chen Ph.D. Assistant Professor Department of Computer Science University of Central Arkansas Fall 2010

2. Outline • Hierarchical Clustering • Hybrid Hierarchical Kmeans clustering • DBscan

3. Hierarchical Clustering Venn Diagram of Clustered Data Dendrogram From http://www.stat.unc.edu/postscript/papers/marron/Stat321FDA/RimaIzempresentation.ppt

4. Nearest Neighbor, Level 2, k = 7 clusters. From http://www.stat.unc.edu/postscript/papers/marron/Stat321FDA/RimaIzempresentation.ppt

5. Nearest Neighbor, Level 3, k = 6 clusters.

6. Nearest Neighbor, Level 4, k = 5 clusters.

7. Nearest Neighbor, Level 5, k = 4 clusters.

8. Nearest Neighbor, Level 6, k = 3 clusters.

9. Nearest Neighbor, Level 7, k = 2 clusters.

10. Nearest Neighbor, Level 8, k = 1 cluster.

11. Functional significant gene clusters Two-way clustering Sample clusters Gene clusters

12. Outline • Hierarchical Clustering • Hybrid Hierarchical Kmeans clustering • DBscan

13. Motivation • Among clustering algorithms, Hierarchical and K-means clustering are the two most popular and classic methods. However, both have their innate disadvantages. • K-means clustering requires a specified number of clusters in advance and chooses initial centroids randomly; in other words, you don’t know how to start • Hierarchical clustering is hard to find a place to cut

14. Hybrid Hierarchical K-means Clustering (HHK) Algorithm • The brief idea is we cluster around half data through Hierarchical clustering and succeed by K-means for the remaining • In order to generate super-rules, we let Hierarchical terminate when it generates the largest number of clusters

15. Hybrid Hierarchical K-means Clustering (HHK) Algorithm

16. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

17. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

18. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

19. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

20. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

21. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

22. Hybrid Hierarchical K-means Clustering (HHK) Algorithm Example

23. Outline • Hierarchical Clustering • Hybrid Hierarchical Kmeans clustering • DBscan

24. Density-Based Clustering Methods • Clustering based on density (local cluster criterion), such as density-connected points • Major features: • Discover clusters of arbitrary shape • Handle noise • One scan • Need density parameters as termination condition

25. DBscan • Two parameters: • Eps: Maximum radius of the neighbourhood • MinPts: Minimum number of points in an Eps-neighbourhood of that point

26. DBscan • Directly density-reachable: A point p is directly density-reachable from a point q w.r.t. Eps, MinPts if • p belongs to NEps(q) • core point condition: |NEps (q)| >= MinPts

27. Outlier Border Eps = 1cm MinPts = 5 Core DBSCAN: Density Based Spatial Clustering of Applications with Noise • Relies on a density-based notion of cluster: A cluster is defined as a maximal set of density-connected points • Discovers clusters of arbitrary shape in spatial databases with noise

28. DBscan • Arbitrary select a point p • Retrieve all points density-reachable from p w.r.t. Eps and MinPts. • If p is a core point, a cluster is formed. • If p is a border point, no points are density-reachable from p and DBSCAN visits the next point of the database. • Continue the process until all of the points have been processed.

29. DBSCAN: Sensitive to Parameters