Clustering is the problem of partitioning data points into groups based on their similarity. Correlation clustering provides a method for clustering a set of objects into the optimum number of clusters without specifying that number in advance.[1]
G=(V,E)
k
It may not be possible to find a perfect clustering, where all similar items are in a cluster while all dissimilar ones are in different clusters. If the graph indeed admits a perfect clustering, then simply deleting all the negative edges and finding the connected components in the remaining graph will return the required clusters.
But, in general a graph may not have a perfect clustering. For example, given nodes a,b,c such that a,b and a,c are similar while b,c are dissimilar, a perfect clustering is not possible. In such cases, the task is to find a clustering that maximizes the number of agreements (number of + edges inside clusters plus the number of − edges between clusters) or minimizes the number of disagreements (the number of − edges inside clusters plus the number of + edges between clusters). This problem of maximizing the agreements is NP-complete (multiway cut problem reduces to maximizing weighted agreements and the problem of partitioning into triangles[2] can be reduced to the unweighted version).
Let
G=(V,E)
V
E
G
\Pi=\{\pi1,...,\pik\}
V=\pi1\cup...\cup\pik
\pii\cap\pij=\emptyset
i ≠ j
\Pi
\delta(\Pi)=\{\{u,v\}\inE\mid\{u,v\}\not\subseteq\pi \forall\pi\in\Pi\}
G
\Pi
w\colonE\to\R\geq
E=E+\cupE-
E+
E-
The minimum disagreement correlation clustering problem is the following optimization problem:Here, the set
E+\cap\delta(\Pi)
\Pi
E-\setminus\delta(\Pi)
\Pi
\Pi
Similarly to the minimum disagreement correlation clustering problem, the maximum agreement correlation clustering problem is defined as Here, the set
E+\setminus\delta(\Pi)
\Pi
E-\cap\delta(\Pi)
\Pi
\Pi
Instead of formulating the correlation clustering problem in terms of non-negative edge weights and a partition of the edges into attractive and repulsive edges the problem is also formulated in terms of positive and negative edge costs without partitioning the set of edges explicitly.For given weights
w\colonE\to\R\geq
E=E+\cupE-
e\inE
An edge whose endpoints are in different clusters is said to be cut. The set
\delta(\Pi)
G
The minimum cost multicut problem is the problem of finding a clustering
\Pi
G
Similar to the minimum cost multicut problem, coalition structure generation in weighted graph games[4] is the problem of finding a clustering such that the sum of the costs of the edges that are not cut is maximal:This formulation is also known as the clique partitioning problem.[5]
It can be shown that all four problems that are formulated above are equivalent.This means that a clustering that is optimal with respect to any of the four objectives is optimal for all of the four objectives.
Bansal et al.[6] discuss the NP-completeness proof and also present both a constant factor approximation algorithm and polynomial-time approximation scheme to find the clusters in this setting. Ailon et al.[7] propose a randomized 3-approximation algorithm for the same problem.
CC-Pivot(G=(V,E+,E−)) Pick random pivot i ∈ V Set
C=\{i\}
The authors show that the above algorithm is a 3-approximation algorithm for correlation clustering. The best polynomial-time approximation algorithm known at the moment for this problem achieves a ~2.06 approximation by rounding a linear program, as shown by Chawla, Makarychev, Schramm, and Yaroslavtsev.[8]
Karpinski and Schudy[9] proved existence of a polynomial time approximation scheme (PTAS) for that problem on complete graphs and fixed number of clusters.
In 2011, it was shown by Bagon and Galun[10] that the optimization of the correlation clustering functional is closely related to well known discrete optimization methods.In their work they proposed a probabilistic analysis of the underlying implicit model that allows the correlation clustering functional to estimate the underlying number of clusters.This analysis suggests the functional assumes a uniform prior over all possible partitions regardless of their number of clusters.Thus, a non-uniform prior over the number of clusters emerges.
Several discrete optimization algorithms are proposed in this work that scales gracefully with the number of elements (experiments show results with more than 100,000 variables).The work of Bagon and Galun also evaluated the effectiveness of the recovery of the underlying number of clusters in several applications.
Correlation clustering also relates to a different task, where correlations among attributes of feature vectors in a high-dimensional space are assumed to exist guiding the clustering process. These correlations may be different in different clusters, thus a global decorrelation cannot reduce this to traditional (uncorrelated) clustering.
Correlations among subsets of attributes result in different spatial shapes of clusters. Hence, the similarity between cluster objects is defined by taking into account the local correlation patterns. With this notion, the term has been introduced in [11] simultaneously with the notion discussed above.Different methods for correlation clustering of this type are discussed in [12] and the relationship to different types of clustering is discussed in.[13] See also Clustering high-dimensional data.
Correlation clustering (according to this definition) can be shown to be closely related to biclustering. As in biclustering, the goal is to identify groups of objects that share a correlation in some of their attributes; where the correlation is usually typical for the individual clusters.