CS570: Introduction to Data Mining

Transcription

1 CS570: Introduction to Data Mining Scalable Clustering Methods: BIRCH and Others Reading: Chapter 10.3 Han, Chapter 9.5 Tan Cengiz Gunay, Ph.D. Slides courtesy of Li Xiong, Ph.D., 2011 Han, Kamber & Pei. Data Mining. Morgan Kaufmann, and 2006 Tan, Steinbach & Kumar. Introd. Data Mining., Pearson. Addison Wesley. October 7,

2 Previously: Hierarchical Clustering Produces a set of nested clusters organized as a hierarchical tree Can be visualized as a dendrogram, a tree like diagram Clustering obtained by cutting at desired level Do not have to assume any particular number of clusters May correspond to meaningful taxonomies October 7,

3 Previously: Major Weaknesses of Hierarchical Clustering Do not scale well (N: number of points) Space complexity: Time complexity: O(N 2 ) O(N 3 ) O(N 2 log(n)) for some cases/approaches Cannot undo what was done previously Quality varies in terms of distance measures MIN (single link): susceptible to noise/outliers MAX/GROUP AVERAGE: may not work well with nonglobular clusters How to improve?

4 Scalable Hierarchical Clustering Methods Combines hierarchical and partitioning approaches Recent methods: BIRCH (1996): uses CF-tree and incrementally adjusts the quality of sub-clusters CURE(1998): uses representative points for inter-cluster distance ROCK (1999): clustering categorical data by neighbor and link analysis CHAMELEON (1999): hierarchical clustering using dynamic modeling on graphs October 7, 2013 Data Mining: Concepts and Techniques 4

5 BIRCH A Tree-based Approach October 8, 2013 Data Mining: Scalable Clustering 5

6 BIRCH BIRCH: Balanced Iterative Reducing and Clustering using Hierarchies (Zhang, Ramakrishnan & Livny, SIGMOD 96) SIGMOD 10-year test of time award Main ideas: Incremental (does not need the whole dataset) Summarizes using in-memory clustering feature Combines hierarchical clustering for microclustering and partitioning for macroclustering Features: Scales linearly: single scan and improves the quality with a few additional scans Weakness: handles only numeric data, and sensitive to the order of the data records. October 8, 2013 Data Mining: Concepts and Techniques 6

7 Cluster Statistics Given a cluster of instances Centroid: Radius: average distance from member points to centroid Diameter: average pair-wise distance within a cluster October 7,

8 Given two clusters Intra-Cluster Distance Centroid Euclidean distance: Centroid Manhattan distance: Average distance: October 7,

9 Clustering Feature (CF) in BIRCH CF = (5, (16,30),(54,190)) What is it good for? October 7, (3,4) (2,6) (4,5) (4,7) (3,8)

10 Properties of Clustering Feature CF entry is more compact Stores significantly less then all of the data points in the sub-cluster A CF entry has sufficient information to calculate statistics about the cluster and intra-cluster distances Additivity theorem allows us to merge subclusters incrementally & consistently October 7,

11 Cluster Statistics from CF CF: n, LS, SS D Radius (average distance to centroid) R n i x x / n nss 2LS nls n i 0 / Diameter (average pairwise distance) n n 2 2 x x / n n 1 2nSS 2LS / n n 1 i j i 1 j 1 October 7, 2013 Data Mining: Concepts and Techniques 11

12 Hierarchical CF-Tree A CF tree is a height-balanced tree that stores the clustering features for a hierarchical clustering A nonleaf node in a tree has descendants or children The nonleaf nodes store sums of the CFs of their children A CF tree has two parameters Branching factor: maximum number of children. Threshold: max diameter of sub-clusters stored at the leaf nodes Why a tree? October 7, 2013 Data Mining: Concepts and Techniques 12

13 CF 1 The CF Tree Structure CF 2 Root child 1 child 2 child 3 child 6 CF 3 CF 6 CF 1 CF 2 Non-leaf node child 1 child 2 child 3 child 5 CF 3 CF 5 Leaf node Leaf node prev CF 1 CF 2 CF 6 next prev CF 1 CF 2 CF 4 next October 7, 2013 Data Mining: Concepts and Techniques 13

14 CF-Tree Insertion Traverse down from root (top-down), find the appropriate leaf Follow the "closest"-cf path, w.r.t. intracluster distance measures Modify the leaf If the closest-cf leaf cannot absorb, make a new CF entry. If there is no room for new leaf, split the parent node Traverse back & up Updating CFs on the path or splitting nodes October 7,

15 BIRCH Overview October 7,

16 The Algorithm: BIRCH Phase 1: Scan database to build an initial inmemory CF-tree Subsequent phases become fast, accurate, less order sensitive Phase 2: Condense data (optional) Rebuild the CF-tree with a larger T (why?) Phase 3: Global clustering Use existing clustering algorithm on CF entries Helps fix problem where natural clusters span nodes Phase 4: Cluster refining (optional) Do additional passes over the dataset & reassign data points to the closest centroid from phase 3 October 7,

17 Main ideas: BIRCH Summary Incremental (does not need the whole dataset) Use in-memory clustering feature to summarize Use hierarchical clustering for microclustering and other clustering methods (e.g. partitioning) for macroclustering Features: Scales linearly: single scan and improves the quality with a few additional scans Weakness: handles only numeric data sensitive to the order of the data records unnatural clusters because of leaf node limit spherical clusters because of diameter (how to solve?) October 7, 2013 Data Mining: Concepts and Techniques 17

18 CURE CURE: Clustering Using REpresentatives CURE: An Efficient Clustering Algorithm for Large Databases (1998) S Guha, R Rastogi, K Shim Addresses potential problems with min, max, centroid distance based hierarchical clustering Main ideas: Use representative points for inter-cluster distance Random sampling and partitioning Features: More robust to outliers Better for non-spherical shapes and non-uniform sizes See Tan Ch 9.5.3, pp

19 CURE: Cluster Points A number of points (e.g., 10+) represents a cluster Representative points: Start with farthest from centroid Add farthest from all selected, and so on until k points Shrink them by distance to center of the cluster Why shrink? Parallels to other methods? Cluster similarity is the similarity of the closest pair of representative points from different clusters

20 Experimental Results: CURE Picture from CURE, Guha, Rastogi, Shim.

21 Experimental Results: CURE (centroid) (single link) Picture from CURE, Guha, Rastogi, Shim.

22 CURE Cannot Handle Differing Densities Original Points CURE So far only numerical data?

23 Clustering Categorical Data: The ROCK Algorithm ROCK: RObust Clustering using links S. Guha, R. Rastogi & K. Shim, Int. Conf. Data Eng. (ICDE) 99 Major ideas Use links to measure similarity/proximity Sampling-based clustering Features: More meaningful clusters Emphasizes interconnectivity but ignores proximity Not in textbook, see paper above. October 7, 2013 Data Mining: Concepts and Techniques 23

24 Similarity Measure in ROCK Market basket data clustering Jaccard coefficient-based similarity function: Sim( T, T ) 1 2 T T T 1 2 T 1 2 Example: Two groups (clusters) of transactions C 1. <a, b, c, d, e> {a, b, c}, {a, b, d}, {a, b, e}, {a, c, d}, {a, c, e}, {a, d, e}, {b, c, d}, {b, c, e}, {b, d, e}, {c, d, e} C 2. <a, b, f, g> {a, b, f}, {a, b, g}, {a, f, g}, {b, f, g} Let T 1 = {a, b, c}, T 2 = {c, d, e}, T 3 = {a, b, f} Jaccard coefficient may lead to wrong clustering result: { c} 1 Sim( T 1, T 2) { a, b, c, d, e} 5 { c, f } 2 Sim( T, T3 ) 1 { a, b, c, f } 4 October 7, 2013 Data Mining: Scalable Clustering

25 Link Measure in ROCK Neighbor: Sim( i, j) Links: # of common neighbors Reminder: Let T 1 = {a, b, c}, T 2 = {c, d, e}, T 3 = {a, b, f} Let C 1 : <a, b, c, d, e>, C 2 : <a, b, f, g> Example: link(t 1, T 2 ) = 4, since they have 4 common neighbors {a, c, d}, {a, c, e}, {b, c, d}, {b, c, e} link(t 1, T 3 ) = 3, since they have 3 common neighbors {a, b, d}, {a, b, e}, {a, b, g} October 7, 2013 Data Mining: Concepts and Techniques 25

26 Rock Algorithm 1. Obtain a sample of points from the data set 2. Compute the link value for each set of points (computed by Jaccard coefficient) 3. Perform an agglomerative (bottom-up) hierarchical clustering on the data using the number of shared neighbors as similarity measure 4. Assign the remaining points to the clusters that have been found October 7, 2013 Data Mining: Concepts and Techniques 26

27 Cluster Merging: Limitations of Current Schemes Existing schemes are static in nature: MIN or CURE merge two clusters based on their closeness (or minimum distance) GROUP-AVERAGE or ROCK: merge two clusters based on their average connectivity

28 Limitations of Current Merging Schemes (a) (b) (c) (d) Closeness schemes will merge (a) and (b) Solution? Average connectivity schemes will merge (c) and (d)

29 Chameleon: Clustering Using Dynamic Modeling Adapt to the characteristics of the data set to find the natural clusters Use a dynamic model to measure the similarity between clusters Main property is the relative closeness and relative interconnectivity of the cluster Two clusters are combined if the resulting cluster shares certain properties with the constituent clusters The merging scheme preserves self-similarity

30 CHAMELEON: Hierarchical Clustering Using Dynamic Modeling (1999) CHAMELEON: by G. Karypis, E.H. Han, and V. Kumar 99 Basic ideas: A graph-based clustering approach A two-phase algorithm: Partitioning: cluster objects into a large number of small sub-clusters Agglomerative hierarchical clustering: repeatedly combine sub-clusters Measures the similarity based on a dynamic model Features: interconnectivity and closeness (proximity) Handles clusters of arbitrary shapes, sizes, and density Scales well See Sections Han and Tan October 7, 2013 Data Mining: Concepts and Techniques 30

31 Graph-Based Clustering Uses the proximity graph Start with the proximity matrix Consider each point as a node in a graph Each edge between two nodes has a weight which is the proximity between the two points Fully connected proximity graph MIN (single-link) and MAX (complete-link) Sparsification eliminates data for graph algos. knn Threshold based Clusters are connected components in the graph

32 Overall Framework of CHAMELEON Construct Sparse Graph Partition the Graph Data Set Merge Partition Final Clusters October 7, 2013 Data Mining: Concepts and Techniques 32

33 Chameleon: Steps Preprocessing Step: Represent the Data by a Graph Given a set of points, construct the k-nearest-neighbor (k-nn) graph to capture the relationship between a point and its k nearest neighbors Concept of neighborhood is captured dynamically (even if region is sparse) Phase 1: Use a multilevel graph partitioning algorithm on the graph to find a large number of clusters of wellconnected vertices Each cluster should contain mostly points from one true cluster, i.e., is a sub-cluster of a real cluster

34 Chameleon: Steps (cont) Phase 2: Use Hierarchical Agglomerative (bottom-up) Clustering to merge sub-clusters Two clusters are combined if the resulting cluster shares certain properties with the constituent clusters Two key properties used to model cluster similarity: Relative Interconnectivity: Absolute interconnectivity of two clusters normalized by the internal connectivity of the clusters Relative Closeness: Absolute closeness of two clusters normalized by the internal closeness of the clusters

35 CHAMELEON (Clustering Complex Objects) October 7, 2013 Data Mining: Concepts and Techniques 35

36 Chapter 7. Cluster Analysis Overview Partitioning methods Hierarchical methods Density-based methods Other methods Cluster evaluation Outlier analysis Summary October 7, 2013 Data Mining: Concepts and Techniques 36