Cluster analysis


Jump to:navigation, search

Cluster analysis or clustering is the assignment of a set of observations into subsets (called clusters) so that observations in the same cluster are similar in some sense. Clustering is a method of unsupervised learning, and a common technique for statistical data analysis used in many fields, including machine learning, data mining, pattern recognition, image analysis and bioinformatics.

Besides the term clustering, there are a number of terms with similar meanings, including automatic classification, numerical taxonomy, botryology and typological analysis.


[edit] Types of clustering

Data clustering algorithms can be hierarchical. Hierarchical algorithms find successive clusters using previously established clusters. These algorithms can be either agglomerative ("bottom-up") or divisive ("top-down"). Agglomerative algorithms begin with each element as a separate cluster and merge them into successively larger clusters. Divisive algorithms begin with the whole set and proceed to divide it into successively smaller clusters.

Partitional algorithms typically determine all clusters at once, but can also be used as divisive algorithms in the hierarchical clustering.

Density-based clustering algorithms are devised to discover arbitrary-shaped clusters. In this approach, a cluster is regarded as a region in which the density of data objects exceeds a threshold. DBSCAN and OPTICS are two typical algorithms of this kind.

Two-way clustering, co-clustering or biclustering are clustering methods where not only the objects are clustered but also the features of the objects, i.e., if the data is represented in a data matrix, the rows and columns are clustered simultaneously.

Another important distinction is whether the clustering uses symmetric or asymmetric distances. A property of Euclidean space is that distances are symmetric (the distance from object A to B is the same as the distance from B to A). In other applications (e.g., sequence-alignment methods, see Prinzie & Van den Poel (2006)), this is not the case.

Many clustering algorithms require specification of the number of clusters to produce in the input data set, prior to execution of the algorithm. Barring knowledge of the proper value beforehand, the appropriate value must be determined, a problem for which a number of techniques have been developed.

[edit] Distance measure

An important step in any clustering is to select a distance measure, which will determine how the similarity of two elements is calculated. This will influence the shape of the clusters, as some elements may be close to one another according to one distance and farther away according to another. For example, in a 2-dimensional space, the distance between the point (x = 1, y = 0) and the origin (x = 0, y = 0) is always 1 according to the usual norms, but the distance between the point (x = 1, y = 1) and the origin can be 2, √2 or 1 if you take respectively the 1-norm, 2-norm or infinity-norm distance.

Common distance functions:

[edit] Hierarchical clustering

Hierarchical clustering creates a hierarchy of clusters which may be represented in a tree structure called a dendrogram. The root of the tree consists of a single cluster containing all observations, and the leaves correspond to individual observations.

Algorithms for hierarchical clustering are generally either agglomerative, in which one starts at the leaves and successively merges clusters together; or divisive, in which one starts at the root and recursively splits the clusters.

Any valid metric may be used as a measure of similarity between pairs of observations. The choice of which clusters to merge or split is determined by a linkage criteria, which is a function of the pairwise distances between observations.

Cutting the tree at a given height will give a clustering at a selected precision. In the following example, cutting after the second row will yield clusters {a} {b c} {d e} {f}. Cutting after the third row will yield clusters {a} {b c} {d e f}, which is a coarser clustering, with a smaller number of larger clusters.

[edit] Agglomerative hierarchical clustering

For example, suppose this data is to be clustered, and the euclidean distance is the distance metric.

Raw data

The hierarchical clustering dendrogram would be as such:

Traditional representation

This method builds the hierarchy from the individual elements by progressively merging clusters. In our example, we have six elements {a} {b} {c} {d} {e} and {f}. The first step is to determine which elements to merge in a cluster. Usually, we want to take the two closest elements, according to the chosen distance.

Optionally, one can also construct a distance matrix at this stage, where the number in the i-th row j-th column is the distance between the i-th and j-th elements. Then, as clustering progresses, rows and columns are merged as the clusters are merged and the distances updated. This is a common way to implement this type of clustering, and has the benefit of caching distances between clusters. A simple agglomerative clustering algorithm is described in the single-linkage clustering page; it can easily be adapted to different types of linkage (see below).

Suppose we have merged the two closest elements b and c, we now have the following clusters {a}, {b, c}, {d}, {e} and {f}, and want to merge them further. To do that, we need to take the distance between {a} and {b c}, and therefore define the distance between two clusters. Usually the distance between two clusters \mathcal{A} and \mathcal{B} is one of the following:

 \max \{\, d(x,y) : x \in \mathcal{A},\, y \in \mathcal{B}\,\}.
 \min \{\, d(x,y) : x \in \mathcal{A},\, y \in \mathcal{B} \,\}.
 {1 \over {|\mathcal{A}|\cdot|\mathcal{B}|}}\sum_{x \in \mathcal{A}}\sum_{ y \in \mathcal{B}} d(x,y).

Each agglomeration occurs at a greater distance between clusters than the previous agglomeration, and one can decide to stop clustering either when the clusters are too far apart to be merged (distance criterion) or when there is a sufficiently small number of clusters (number criterion).

[edit] Concept clustering

Another variation of the agglomerative clustering approach is conceptual clustering.

[edit] Partitional clustering

[edit] K-means and derivatives

[edit] k-means clustering

The k-means algorithm assigns each point to the cluster whose center (also called centroid) is nearest. The center is the average of all the points in the cluster — that is, its coordinates are the arithmetic mean for each dimension separately over all the points in the cluster.

Example: The data set has three dimensions and the cluster has two points: X=(x_1,x_2,x_3)\, and Y=(y_1,y_2,y_3)\,. Then the centroid Z becomes Z=(z_1,z_2,z_3)\,, where:

z_1=\frac{x_1+y_1}{2}, z_2=\frac{x_2+y_2}{2} and z_3=\frac{x_3+y_3}{2}

The algorithm steps are [1]:

The main advantages of this algorithm are its simplicity and speed which allows it to run on large datasets. Its disadvantage is that it does not yield the same result with each run, since the resulting clusters depend on the initial random assignments. It minimizes intra-cluster variance, but does not ensure that the result has a global minimum of variance. Another disadvantage is the requirement for the concept of a mean to be definable which is not always the case. For such datasets the k-medoids variant is appropriate.

[edit] Fuzzy c-means clustering

In fuzzy clustering, each point has a degree of belonging to clusters, as in fuzzy logic, rather than belonging completely to just one cluster. Thus, points on the edge of a cluster, may be in the cluster to a lesser degree than points in the center of cluster. For each point x we have a coefficient giving the degree of being in the kth cluster uk(k). Usually, the sum of those coefficients for any given x is defined to be 1:

 \forall x \left(\sum_{k=1}^{\mathrm{num.}\ \mathrm{clusters}} u_k(x) \ =1\right)

With fuzzy c-means, the centroid of a cluster is the mean of all points, weighted by their degree of belonging to the cluster:

\mathrm{center}_k = {{\sum_x u_k(x)^m x} \over {\sum_x u_k(x)^m}}

The degree of belonging is related to the inverse of the distance to the cluster center:

u_k(x) = {1 \over d(\mathrm{center}_k,x)},

then the coefficients are normalized and fuzzyfied with a real parameter m > 1 so that their sum is 1. So

u_k(x) = \frac{1}{\sum_j \left(\frac{d(\mathrm{center}_k,x)}{d(\mathrm{center}_j,x)}\right)^{2/(m-1)}}

For m equal to 2, this is equivalent to normalising the coefficient linearly to make their sum 1. When m is close to 1, then cluster center closest to the point is given much more weight than the others, and the algorithm is similar to k-means.

The fuzzy c-means algorithm is very similar to the k-means algorithm: [2]

The algorithm minimizes intra-cluster variance as well, but has the same problems as k-means, the minimum is a local minimum, and the results depend on the initial choice of weights. The expectation-maximization algorithm is a more statistically formalized method which includes some of these ideas: partial membership in classes. It has better convergence properties and is in general preferred to fuzzy-c-means.

[edit] QT clustering algorithm

QT (quality threshold) clustering (Heyer, Kruglyak, Yooseph, 1999) is an alternative method of partitioning data, invented for gene clustering. It requires more computing power than k-means, but does not require specifying the number of clusters a priori, and always returns the same result when run several times.

The algorithm is:

The distance between a point and a group of points is computed using complete linkage, i.e. as the maximum distance from the point to any member of the group (see the "Agglomerative hierarchical clustering" section about distance between clusters).

[edit] Locality-sensitive hashing

Locality-sensitive hashing can be used for clustering. Feature space vectors are sets, and the metric used is the Jaccard distance. The feature space can be considered high-dimensional. The min-wise independent permutations LSH scheme (sometimes MinHash) is then used to put similar items into buckets. With just one set of hashing methods, there are only clusters of very similar elements. By seeding the hash functions several times (eg 20), it is possible to get bigger clusters. [3]

[edit] Graph-theoretic methods

Formal concept analysis is a technique for generating clusters of objects and attributes, given a bipartite graph representing the relations between the objects and attributes. Other methods for generating overlapping clusters (a cover rather than a partition) are discussed by Jardine and Sibson (1968) and Cole and Wishart (1970).

[edit] Spectral clustering

See also: Kernel principal component analysis

Given a set of data points A, the similarity matrix may be defined as a matrix S where Sij represents a measure of the similarity between points i, j\in A. Spectral clustering techniques make use of the spectrum of the similarity matrix of the data to perform dimensionality reduction for clustering in fewer dimensions.

One such technique is the Shi-Malik algorithm, commonly used for image segmentation. It partitions points into two sets (S1,S2) based on the eigenvector v corresponding to the second-smallest eigenvalue of the Laplacian matrix

L = I - D^{-1/2}SD^{-1/2}\;

of S, where D is the diagonal matrix

D_{ii} = \sum_{j} S_{ij}\;

This partitioning may be done in various ways, such as by taking the median m of the components in v, and placing all points whose component in v is greater than m in S1, and the rest in S2. The algorithm can be used for hierarchical clustering by repeatedly partitioning the subsets in this fashion.

A related algorithm is the Meila-Shi algorithm, which takes the eigenvectors corresponding to the k largest eigenvalues of the matrix P = SD − 1 for some k, and then invokes another (e.g. k-means) to cluster points by their respective k components in these eigenvectors.

[edit] Applications

[edit] Biology

In biology clustering has many applications

[edit] Medicine

In medical imaging, such as PET scans, cluster analysis can be used to differentiate between different types of tissue and blood in a three dimensional image. In this application, actual position does not matter, but the voxel intensity is considered as a vector, with a dimension for each image that was taken over time. This technique allows, for example, accurate measurement of the rate a radioactive tracer is delivered to the area of interest, without a separate sampling of arterial blood, an intrusive technique that is most common today.

[edit] Market research

Cluster analysis is widely used in market research when working with multivariate data from surveys and test panels. Market researchers use cluster analysis to partition the general population of consumers into market segments and to better understand the relationships between different groups of consumers/potential customers.

[edit] Other applications

Social network analysis
In the study of social networks, clustering may be used to recognize communities within large groups of people.
Image segmentation
Clustering can be used to divide a digital image into distinct regions for border detection or object recognition.
Data mining
Many data mining applications involve partitioning data items into related subsets; the marketing applications discussed above represent some examples. Another common application is the division of documents, such as World Wide Web pages, into genres.
Search result grouping
In the process of intelligent grouping of the files and websites, clustering may be used to create a more relevant set of search results compared to normal search engines like Google. There are currently a number of web based clustering tools such as Clusty.
Slippy map optimization
Flickr's map of photos and other map sites use clustering to reduce the number of markers on a map. This makes it both faster and reduces the amount of visual clutter.
IMRT segmentation
Clustering can be used to divide a fluence map into distinct regions for conversion into deliverable fields in MLC-based Radiation Therapy.
Grouping of Shopping Items
Clustering can be used to group all the shopping items available on the web into a set of unique products. For example, all the items on eBay can be grouped into unique products. (eBay doesn't have the concept of a SKU)
Mathematical chemistry
To find structural similarity, etc., for example, 3000 chemical compounds were clustered in the space of 90 topological indices.[4]
Petroleum Geology
Cluster Analysis is used to reconstruct missing bottom hole core data or missing log curves in order to evaluate reservoir properties.
Physical Geography
The clustering of chemical properties in different sample locations.

[edit] Comparisons between data clusterings

There have been several suggestions for a measure of similarity between two clusterings. Such a measure can be used to compare how well different data clustering algorithms perform on a set of data. Many of these measures are derived from the matching matrix (aka confusion matrix), e.g., the Rand measure, Adjusted Rand Index and the Fowlkes-Mallows Bk measures.[5] With small variations, such kind of measures also include set matching based clustering criteria, e.g., the Meila and Heckerman.

Several different clustering systems based on mutual information have been proposed. One is Marina Meila's 'Variation of Information' metric (see ref below); another provides hierarchical clustering[6]. The adjusted-for-chance version for the mutual information is the Adjusted Mutual Information (AMI), which corrects mutual information for agreement solely due to chance between clusterings (Vinh et. al, 2009).

Recently, a new Mallows Distance based metric was also proposed for soft clustering comparisons.[citation needed]

[edit] Algorithms

In recent years considerable effort has been put into improving algorithm performance (Z. Huang, 1998). Among the most popular are CLARANS (Ng and Han,1994), DBSCAN (Ester et al., 1996) and BIRCH (Zhang et al., 1996).

[edit] See also

[edit] Bibliography

  1. J. MacQueen, 1967
  2. Bezdek, James C. (1981), Pattern Recognition with Fuzzy Objective Function Algorithms, ISBN 0306406713 
  3. Google News personalization: scalable online collaborative filtering
  4. Basak S.C., Magnuson V.R., Niemi C.J., Regal R.R. "Determing Structural Similarity of Chemicals Using Graph Theoretic Indices". Discr. Appl. Math., 19, 1988: 17-44.
  5. E. B. Fowlkes & C. L. Mallows (September 1983). "A Method for Comparing Two Hierarchical Clusterings". Journal of the American Statistical Association 78 (383): 553–584. doi:10.2307/2288117. 
  6. Alexander Kraskov, Harald Stögbauer, Ralph G. Andrzejak, and Peter Grassberger, "Hierarchical Clustering Based on Mutual Information", (2003) ArXiv q-bio/0311039

[edit] Others

For spectral clustering:

For estimating number of clusters:

For discussion of the elbow criterion:

[edit] External links

Need Help
Share This Page