lecture 13: clustering (continued) may 12, 2010
DESCRIPTION
Lecture 13: Clustering (continued) May 12, 2010. Announcements end of next class (May 19), at 8:45, take-home finals will be given. Due: May 26 at 6 PM project presentation – more in the next slide project report due: May 26 at the time of presentation. - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/1.jpg)
Lecture 13: Clustering (continued) May 12, 2010
![Page 2: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/2.jpg)
Announcements
• end of next class (May 19), at 8:45, take-home finals will be given.
• Due: May 26 at 6 PM
• project presentation – more in the next slide
• project report due: May 26 at the time of presentation.
![Page 3: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/3.jpg)
Project Presentation – Details• recommended format: slides with overhead
projector (e.g. power-point)
• sample presentations – e-mailed and can be found under project link
• duration: 15 minutes
• should include:• Problem statement• Data set – size, how acquired, processing needed• Algorithm – overview, time and space needs• Result – performance, plots• Challenges • Summary and conclusion
![Page 4: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/4.jpg)
K-Means
Assumes documents are real-valued vectors. Clusters based on centroids (aka the center of gravity or mean) of points in a cluster, c:
Reassignment of instances to clusters is based on distance to the current cluster centroids.
cx
xc
||
1(c)μ
Sec. 16.4
![Page 5: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/5.jpg)
K-Means Algorithm
Select K random docs {s1, s2,… sK} as seeds.Until clustering converges (or other stopping criterion): for each doc di: Assign di to the cluster cj such that dist(xi, sj) is minimal.
(Next, update the seeds to the centroid of each cluster) for each cluster cj
sj = (cj)
Sec. 16.4
![Page 6: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/6.jpg)
More formal description of algorithm
6
![Page 7: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/7.jpg)
K Means Example(K=2)
Pick seeds
Reassign clusters
Compute centroids
xx
Reassign clusters
xx xx Compute centroids
Reassign clusters
Converged!
Sec. 16.4
![Page 8: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/8.jpg)
Termination conditions
Several possibilities, e.g., A fixed number of iterations. Doc partition unchanged. Centroid positions don’t change.
Does this mean that the docs in a cluster are
unchanged?
Sec. 16.4
![Page 9: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/9.jpg)
Convergence
Why should the K-means algorithm ever reach a fixed point? A state in which clusters don’t change.
K-means is a special case of a general procedure known as the Expectation Maximization (EM) algorithm. EM is known to converge. Number of iterations could be large.
But in practice usually isn’t
Sec. 16.4
![Page 10: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/10.jpg)
Convergence of K-Means
Define goodness measure of cluster k as sum of squared distances from cluster centroid: Gk = Σi (di – ck)2 (sum over all di in cluster k)
G = Σk Gk
Reassignment monotonically decreases G since each vector is assigned to the closest centroid.
Sec. 16.4
![Page 11: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/11.jpg)
Convergence of K-Means
Recomputation monotonically decreases each Gk since (mk is number of members in cluster k):
Σ (di – a)2 reaches minimum for:
Σ –2(di – a) = 0
Σ di = Σ a
mK a = Σ di
a = (1/ mk) Σ di = ck
K-means typically converges quickly
Sec. 16.4
![Page 12: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/12.jpg)
Time Complexity
Computing distance between two docs is O(M) where M is the dimensionality of the vectors.
Reassigning clusters: O(KN) distance computations, or O(KNM).
Computing centroids: Each doc gets added once to some centroid: O(NM).
Assume these two steps are each done once for I iterations.
Total time = O(IKNM).However, it is not clear how to bound I unless it is forced externally.
Sec. 16.4
![Page 13: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/13.jpg)
Seed Choice Results can vary based on random seed selection.
Some seeds can result in poor convergence rate, or convergence to sub-optimal clusters. Select good seeds using a heuristic (e.g., doc least similar to any existing mean)
Try out multiple starting points
Initialize with the results of another method.
In the above, if you startwith B and E as centroidsyou converge to {A,B,C}and {D,E,F}If you start with D and Fyou converge to {A,B,D,E} {C,F}
Example showingsensitivity to seeds
Sec. 16.4
![Page 14: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/14.jpg)
Two different K-means Clusterings
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Sub-optimal Clustering
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Optimal Clustering
Original Points
![Page 15: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/15.jpg)
Problem with Selecting Initial centroids
If there are K ‘real’ clusters then the chance of selecting one centroid from each cluster is small. Chance is relatively small when K is large If clusters are the same size, n, then
For example, if K = 10, then probability = 10!/1010 = 0.00036 Sometimes the initial centroids will readjust themselves in ‘right’ way, and sometimes they don’t
![Page 16: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/16.jpg)
Initial Centroids not well chosen
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 1
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 3
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 4
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 5
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 6
![Page 17: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/17.jpg)
A seemingly better initial choice
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 1
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 2
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 3
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
x
y
Iteration 4
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2
0
0.5
1
1.5
2
2.5
3
xy
Iteration 5
![Page 18: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/18.jpg)
Solutions to Initial Centroids Problem
Multiple runs Helps, but probability is not on your side
Sample and use hierarchical clustering to determine initial centroids
Select more than k initial centroids and then select among these initial centroids Select most widely separated
Postprocessing Bisecting K-means
Not as susceptible to initialization issues
![Page 19: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/19.jpg)
Evaluating K-means Clusters
Most common measure is Sum of Squared Error (SSE) For each point, the error is the distance to the nearest cluster
To get SSE, we square these errors and sum them.
x is a data point in cluster Ci and mi is the representative point for cluster Ci
can show that mi corresponds to the center (mean) of the cluster
Given two clusters, we can choose the one with the smallest error
One easy way to reduce SSE is to increase K, the number of clusters
A good clustering with smaller K can have a lower SSE than a poor clustering with higher K
K
i Cxi
i
xmdistSSE1
2 ),(
![Page 20: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/20.jpg)
K-means: issues, variations, etc.
Recomputing the centroid after every assignment (rather than after all points are re-assigned) can improve speed of convergence of K-means.
Assumes clusters are spherical in vector space Sensitive to coordinate changes, weighting etc.
Disjoint and exhaustive Doesn’t have a notion of “outliers” by default
But can add outlier filtering
Sec. 16.4
![Page 21: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/21.jpg)
How Many Clusters?
Number of clusters K is given Partition n docs into predetermined number of clusters
Finding the “right” number of clusters is part of the problem Given docs, partition into an “appropriate” number of subsets.
E.g., for query results - ideal value of K not known up front - though UI may impose limits.
Can usually take an algorithm for one flavor and convert to the other.
![Page 22: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/22.jpg)
K not specified in advance
Say, the results of a query. Solve an optimization problem: penalize having lots of clusters application dependent, e.g., compressed summary of search results list.
Tradeoff between having more clusters (better focus within each cluster) and having too many clusters
![Page 23: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/23.jpg)
K not specified in advance
Given a clustering, define the benefit for a doc to be the cosine similarity to its centroid.
Define the total benefit to be the sum of the individual doc benefits.
![Page 24: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/24.jpg)
Penalize lots of clusters
For each cluster, we have a Cost C. Thus for a clustering with K clusters, the Total Cost is KC.
Define the Value of a clustering to be = Total Benefit - Total Cost.
Find the clustering of highest value, over all choices of K. Total benefit increases with increasing K. But can stop when it doesn’t increase by “much”. The Cost term enforces this.
![Page 25: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/25.jpg)
Error as a function of k
25
![Page 26: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/26.jpg)
Pre-processing and Post-processing
Pre-processing Normalize the data Eliminate outliers
Post-processing Eliminate small clusters that may represent outliers
Split ‘loose’ clusters, i.e., clusters with relatively high SSE
Merge clusters that are ‘close’ and that have relatively low SSE
Can use these steps during the clustering process
![Page 27: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/27.jpg)
Limitations of K-means: Non-globular Shapes
Original Points K-means (2 Clusters)
![Page 28: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/28.jpg)
Overcoming K-means Limitations
Original Points K-means Clusters
One solution is to use many clusters.Find parts of clusters, but need to put together.
![Page 29: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/29.jpg)
Overcoming K-means Limitations
Original Points K-means Clusters
![Page 30: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/30.jpg)
Overcoming K-means Limitations
Original Points K-means Clusters
![Page 31: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/31.jpg)
Hierarchical Clustering Build a tree-based hierarchical taxonomy (dendrogram) from a set of documents.
One approach: recursive application of a partitional clustering algorithm.
animal
vertebrate
fish reptile amphib. mammal worm insect crustacean
invertebrate
Ch. 17
![Page 32: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/32.jpg)
Dendrogram: Hierarchical Clustering
Clustering obtained by cutting the dendrogram at a desired level: each connected component forms a cluster.
32
![Page 33: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/33.jpg)
Hierarchical Agglomerative Clustering
Starts with each doc in a separate cluster then repeatedly joins the closest pair of clusters, until there is only one cluster.
The history of merging forms a binary tree or hierarchy.
Sec. 17.1
![Page 34: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/34.jpg)
Closest pair of clusters Many variants to defining closest pair of clusters
Single-link Similarity of the most cosine-similar (single-link)
Complete-link Similarity of the “furthest” points, the least cosine-similar
Centroid Clusters whose centroids (centers of gravity) are the most cosine-similar
Average-link Average cosine between pairs of elements
Sec. 17.2
![Page 35: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/35.jpg)
Single Link Agglomerative Clustering
Use maximum similarity of pairs:
Can result in “straggly” (long and thin) clusters due to chaining effect.
After merging ci and cj, the similarity of the resulting cluster to another cluster, ck, is:
),(max),(,
yxsimccsimji cycx
ji
)),(),,(max()),(( kjkikji ccsimccsimcccsim
Sec. 17.2
![Page 36: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/36.jpg)
Single Link ExampleSec. 17.2
![Page 37: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/37.jpg)
Complete Link
Use minimum similarity of pairs:
Makes “tighter,” spherical clusters that are typically preferable.
After merging ci and cj, the similarity of the resulting cluster to another cluster, ck, is:
),(min),(,
yxsimccsimji cycx
ji
)),(),,(min()),(( kjkikji ccsimccsimcccsim
Ci Cj Ck
Sec. 17.2
![Page 38: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/38.jpg)
Complete Link ExampleSec. 17.2
![Page 39: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/39.jpg)
Simple hierarchical clustering algorithm
![Page 40: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/40.jpg)
Computational Complexity
In the first iteration, all HAC methods need to compute similarity of all pairs of N initial instances, which is O(N2).
In each of the subsequent N2 merging iterations, compute the distance between the most recently created cluster and all other existing clusters.
In order to maintain an overall O(N2) performance, computing similarity to each other cluster must be done in constant time. Often O(N3) if done naively or O(N2 log N) if done more cleverly
Sec. 17.2.1
![Page 41: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/41.jpg)
Efficient hierarchical clustering algorithm
![Page 42: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/42.jpg)
Efficient single-link clustering algorithm
![Page 43: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/43.jpg)
Group Average
Similarity of two clusters = average similarity of all pairs within merged cluster.
Compromise between single and complete link.
Two options: Averaged across all ordered pairs in the merged cluster
Averaged over all pairs between the two original clusters
No clear difference in efficacy
)( :)(
),()1(
1),(
ji jiccx xyccyjiji
ji yxsimcccc
ccsim
Sec. 17.3
![Page 44: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/44.jpg)
Computing Group Average Similarity
Always maintain sum of vectors in each cluster.
Compute similarity of clusters in constant time:
jcx
j xcs
)(
)1||||)(|||(|
|)||(|))()(())()((),(
jiji
jijijiji cccc
cccscscscsccsim
Sec. 17.3
![Page 45: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/45.jpg)
What Is A Good Clustering?
Internal criterion: A good clustering will produce high quality clusters in which: the intra-class (that is, intra-cluster) similarity is high
the inter-class similarity is low The measured quality of a clustering depends on both the document representation and the similarity measure used
Sec. 16.3
![Page 46: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/46.jpg)
External criteria for clustering quality
Quality measured by its ability to discover some or all of the hidden patterns or latent classes in gold standard data
Assesses a clustering with respect to ground truth. (requires labeled data)
Assume documents with C gold standard classes, while our clustering algorithms produce K clusters, ω1, ω2, …, ωK with ni members.
Sec. 16.3
![Page 47: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/47.jpg)
External Evaluation of Cluster Quality
Simple measure: purity, the ratio between the dominant class in the cluster πi and the size of cluster ωi
Biased because having n clusters maximizes purity
Others are entropy of classes in clusters (or mutual information between classes and clusters)
Cjnn
Purity ijji
i )(max1
)(
Sec. 16.3
![Page 48: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/48.jpg)
Cluster I Cluster II
Cluster III
Cluster I: Purity = 1/6 (max(5, 1, 0)) = 5/6
Cluster II: Purity = 1/6 (max(1, 4, 1)) = 4/6
Cluster III: Purity = 1/5 (max(2, 0, 3)) = 3/5
Purity exampleSec. 16.3
![Page 49: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/49.jpg)
Rand Index measures between pair decisions
Number of points
Same Cluster in clustering
Different Clusters in clustering
Same class in ground truth
20 24
Different classes in ground truth
20 72
Sec. 16.3
Here RI = 0.68
![Page 50: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/50.jpg)
Rand index and Cluster F-measure
BA
AP
DCBA
DARI
CA
AR
Compare with standard Precision and Recall:
People also define and use a cluster F-measure, which is probably a better measure.
Sec. 16.3
![Page 51: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/51.jpg)
An application of clustering
51
Example: Color quantization of Images ProblemConvert a 24 bit RGB image into a indexed imagewith a palette of K colors.
Solution• The (r, g, b) values of the pixels are the data points xi
• The (r, g, b) values of the K palette colors are the centroids wk.
• Initialize the wk with the color of random pixels.• Perform one pass of k-means algorithm.• Each cluster is assigned one color.
![Page 52: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/52.jpg)
Image Examples
Mnp: 30, percent 0.05, cluster number 4
Mnp : 20, percent 0.05, cluster number 7
Original pictures segmented pictures
Project by Qifong Xu, Penn
![Page 53: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/53.jpg)
Image examples 2
Mnp: 10, percent 0.05, cluster number: 9
Mnp: 50, percent 0.05, cluster number: 3
Original pictures Segmented pictures
![Page 54: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/54.jpg)
Mnp: 30, Cluster number: 4
Mnp:10, cluster number: 15
Original picture
Effect of cluster size
![Page 55: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/55.jpg)
Image clustering in archeology
Angkor Wat temple
• Angkor Wat contains a gallery of 2000 women. • what facial types are represented in these portraits?• A problem being solved by Prof. Anil Jain of MSU using clustering.
![Page 56: Lecture 13: Clustering (continued) May 12, 2010](https://reader036.vdocuments.us/reader036/viewer/2022062408/56813ad6550346895da311e9/html5/thumbnails/56.jpg)
summary
In clustering, clusters are inferred from the data without human input (unsupervised learning)
There are many ways of influencing the outcome of clustering (with user input): number of clusters, similarity measure, choice of features.
Many applications including text clustering, grouping genes/species, image processing/vision etc.