Download presentation
Presentation is loading. Please wait.
Published byEmily Mantel Modified over 10 years ago
1
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs K-means*: Clustering by Gradual Data Transformation Mikko Malinen and Pasi Fränti Speech and Image Processing Unit School of Computing University of Eastern Finland
2
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs K-means* clustering Gradual transformation of data Model Data Fit the data to a model IntermediateFinal
3
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs K-means clustering Iterate between two steps: 1. Assignment step Assign the points to the nearest centroids 2. Update step Update the location of centroids
4
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs K-means* clustering
5
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Example of clustering (s 2 dataset)
6
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 0% done
7
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 10% done
8
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 20% done
9
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 30% done
10
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 40% done
11
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 50% done
12
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 60% done
13
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 70% done
14
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 80% done
15
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 90% done
16
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 100% done
17
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Empty clusters problem
18
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Initialization Data set transform Empty clusters removal K-means Algorithm total Time Complexity
19
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Time Complexity Fixed k-means Initialization Data set transform Empty clusters removal K-means Algorithm total
20
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs s1 d = 2 n = 5000 k = 15 s2 d = 2 n = 5000 k = 15 s3 d = 2 n = 5000 k = 15 s4 d = 2 n = 5000 k = 15 bridge d = 16 n = 4096 k= 256 missa d = 16 n = 6480 k= 256 house d = 3 n=34000 k=256 thyroid d = 5 n = 215 k = 2 iris d = 4 n = 150 k = 2 wine d = 13 n = 178 k = 3 Datasets
21
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error Datasetk-meansproposedGKMoptimal s11.851.010.89 s21.941.521.33 s31.971.711.69 s41.691.631.57 bridge168.2164.7164.1160.7 missa5.335.155.345.12 house9.889.485.945.86 thyroid6.976.921.52 iris3.70 2.02 wine1.921.900.88
22
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
23
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
24
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
25
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
26
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
27
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
28
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Mean square error vs. number of steps
29
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs All correct: Number of incorrect clusters proposed: 36% k-means: 14%
30
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 1 incorrect: Number of incorrect clusters proposed: 64% k-means: 38%
31
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 2 incorrect: Number of incorrect clusters proposed: 0% k-means: 34%
32
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs 3 incorrect: Number of incorrect clusters proposed: 0% k-means: 10%
33
University of Eastern Finland School of Computing P.O. Box 111 FIN- 80101 Joensuu FINLAND Tel. +358 13 251 7959 fax +358 13 251 7955 www.uef.fi/cs Summary We have presented a clustering method based on gradual transformation of data and k-means. Instead of fitting the model to data, we fit the data to a model. The proposed method gives better mean square error than k-means.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.