Consensus Functions for Cluster Ensembles

Pattern Recognition, 2010

Voting-based consensus clustering refers to a distinct class of consensus methods in which the cluster label mismatch problem is explicitly addressed. The voting problem is defined as the problem of finding the optimal relabeling of a given partition with respect to a reference partition. It is commonly formulated as a weighted bipartite matching problem. In this paper, we present a more general formulation of the voting problem as a regression problem with multiple-response and multiple-input variables. We show that a recently introduced cumulative voting scheme is a special case corresponding to a linear regression method. We use a randomized ensemble generation technique, where an overproduced number of clusters is randomly selected for each ensemble partition. We apply an information theoretic algorithm for extracting the consensus clustering from the aggregated ensemble representation and for estimating the number of clusters. We apply it in conjunction with bipartite matching and cumulative voting. We present empirical evidence showing substantial improvements in clustering accuracy, stability, and estimation of the true number of clusters based on cumulative voting. The improvements are achieved in comparison to consensus algorithms based on bipartite matching, which perform very poorly with the chosen ensemble generation technique, and also to other recent consensus algorithms.

The clustering ensembles contains multiple partitions are divided by different clustering algorithms into a single clustering solutions. Clustering ensembles used for improving robustness, stability, and accuracy of unsupervised classification solutions. The major problem of clustering ensemble is the consensus function. Consensus functions in clustering ensembles including hyperactive graph partition, mutual information, co-association based functions, voting approach and finite machine. The characteristics of clustering ensembles algorithm are computational complexity, robustness, simplicity and accuracy on different datasets in previous techniques.

International Journal of Electrical and Computer Engineering (IJECE), 2018

Data analysis plays a prominent role in interpreting various phenomena. Data mining is the process to hypothesize useful knowledge from the extensive data. Based upon the classical statistical prototypes the data can be exploited beyond the storage and management of the data. Cluster analysis a primary investigation with little or no prior knowledge, consists of research and development across a wide variety of communities. Cluster ensembles are melange of individual solutions obtained from different clusterings to produce final quality clustering which is required in wider applications. The method arises in the perspective of increasing robustness, scalability and accuracy. This paper gives a brief overview of the generation methods and consensus functions included in cluster ensemble. The survey is to analyze the various techniques and cluster ensemble methods.

Proceedings of the 2006 SIAM International Conference on Data Mining, 2006

Cluster ensembles offer a solution to challenges inherent to clustering arising from its ill-posed nature. Cluster ensembles can provide robust and stable solutions by leveraging the consensus across multiple clustering results, while averaging out emergent spurious structures that arise due to the various biases to which each participating algorithm is tuned. In this paper, we address the problem of combining multiple weighted clusters which belong to different subspaces of the input space. We leverage the diversity of the input clusterings in order to generate a consensus partition that is superior to the participating ones. Since we are dealing with weighted clusters, our consensus function makes use of the weight vectors associated with the clusters. The experimental results show that our ensemble technique is capable of producing a partition that is as good as or better than the best individual clustering.

Clustering is fundamental to understand the structure of data. In the past decade the cluster ensemble problem has been introduced, which combines a set of partitions (an ensemble) of the data to obtain a single consensus solution that outperforms all the ensemble members. Although disagreement among ensemble partitions (diversity) has been found to be fundamental for success, the literature has arrived to confusing conclusions: some authors suggest that high diversity is beneficial for the final performance, whereas others have indicated that medium is better. While there are several options to measure the diversity, there is no method to control it. This paper introduces a new ensemble generation strategy and a method to smoothly change the ensemble diversity. Experimental results on three datasets suggest that this is an important step towards a more systematic approach to analyze the impact of the ensemble diversity on the overall consensus performance.

Pattern Recognition Letters, 2010

In recent years, the cluster ensembles have been successfully used to tackle well known drawbacks of individual clustering algorithms. Beyond the expected improvement provided by the averaging effect of many clustering algorithms (clustering committee) aiming at the same goal, some interesting experimental results also show that even committees of completely random partitions may lead to a useful consensus. Another powerful finding in cluster ensemble research is that the blind criterion Averaged Normalized Mutual Information seems to replace actual misclassification ratio, whenever labels are given to actual clusters. In this work, we study what is behind these interesting results and the blind criterion, and we use what we learn from this study to propose a new point of view for analysis and design of clustering committees. The usefulness of this new perspective is illustrated through experimental results.

Proceedings of the International Conference on Knowledge Discovery and Information Retrieval, 2009

In order to combine multiple data partitions into a more robust data partition, several approaches to produce the cluster ensemble and various consensus functions have been proposed. This range of possibilities in the multiple data partitions combination raises a new problem: which of the existing approaches, to produce the cluster ensembles' data partitions and to combine these partitions, best fits a given data set. In this paper, we address the cluster ensemble selection problem. We proposed a new measure to select the best consensus data partition, among a variety of consensus partitions, based on a notion of average cluster consistency between each data partition that belongs to the cluster ensemble and a given consensus partition. We compared the proposed measure with other measures for cluster ensemble selection, using 9 different data sets, and the experimental results shown that the consensus partitions selected by our approach usually were of better quality in comparison with the consensus partitions selected by other measures used in our experiments.

Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004., 2004

Clustering ensembles combine multiple partitions of the given data into a single clustering solution of better quality. Inspired by the success of supervised boosting algorithms, we devise an adaptive scheme for integration of multiple non-independent clusterings. Individual partitions in the ensemble are sequentially generated by clustering specially selected subsamples of the given data set. The sampling probability for each data point dynamically depends on the consistency of its previous assignments in the ensemble. New subsamples are drawn to increasingly focus on the problematic regions of the input feature space. A measure of a data point's clustering consistency is defined to guide this adaptation. An empirical study compares the performance of adaptive and regular clustering ensembles using different consensus functions on a number of data sets. Experimental results demonstrate improved accuracy for some clustering structures.

Progress in Pattern …, 2008

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately.

International Journal of Learning Management Systems, 2013

A new criterion for clusters validation is proposed in the paper and based on the new cluster validation criterion a clustering ensmble framework is proposed. The main idea behind the framework is to extract the most stable clusters in terms of the defined criteria. Employing this new cluster validation criterion, the obtained ensemble is evaluated on some well-known and standard data sets. The empirical studies show promising results for the ensemble obtained using the proposed criterion comparing with the ensemble obtained using the standard clusters validation criterion.