Automatic subspace clustering of high dimensional data for data mining applications

Agrawal, Rakesh; Gehrke, Johannes; Gunopulos, Dimitrios; Raghavan, Prabhakar

doi:10.1145/276304.276314

Cited by 1,569 publications

(693 citation statements)

References 25 publications

Supporting

Mentioning

683

Contrasting

Unclassified

Order By: Relevance

“…Subspace clustering has been studied extensively on scalable methods for mining various kinds of subspace clusters, such as grid-based (Agrawal et al 1998), entropy-based (Cheng et al 1999), and density-based (Baumgartner et al 2004;Bohm et al 2004;Kriegel et al 2005;Assent et al 2007) approaches. The above approaches may suffer from the problem of a sheer size of redundancy results.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Efficient approaches for summarizing subspace clusters into k representatives

Chen

Yang

et al. 2010

Soft Comput

View full text Add to dashboard Cite

A major challenge in subspace clustering is that subspace clustering may generate an explosive number of clusters with high computational complexity, which severely restricts the usage of subspace clustering. The problem gets even worse with the increase of the data's dimensionality. In this paper, we propose to summarize the set of subspace clusters into k representative clusters to alleviate the problem. Typically, subspace clusters can be clustered further into k groups, and the set of representative clusters can be selected from each group. In such a way, only the most representative subspace clusters will be returned to user. Unfortunately, when the size of the set of representative clusters is specified, the problem of finding the optimal set is NP-hard. To solve this problem efficiently, we present two approximate methods: PCoC and HCoC. The greatest advantage of our methods is that we only need a subset of subspace clusters as the input instead of the complete set of subspace clusters. Precisely, only the clusters in low-dimensional subspaces are computed and assembled into representative clusters in high-dimensional subspaces. The approximate results can be found in polynomial time. Our performance study shows both the effectiveness and efficiency of these methods.

show abstract

Section: Related Workmentioning

confidence: 99%

“…Condition 3 is denoted as Density Condition which is also used in CLIQUE (Agrawal et al 1998) for identifying dense units. Conceptually, this definition is similar to those in grid-based clustering and density-based clustering.…”

Section: Problem Statementmentioning

confidence: 99%

Efficient approaches for summarizing subspace clusters into k representatives

Chen

Yang

et al. 2010

Soft Comput

View full text Add to dashboard Cite

show abstract

“…A lot of clustering algorithms have been proposed in the literature [1,2]: BIRCH [42], for example, clusters the data with a single scan of the space, since the assignment is based on some local characteristics, without considering the other clusters. An important issue for the problem we deal with is to discover data clusters that follow a Gaussian distribution in a multi-dimensional space.…”

Section: The Clustering Proceduresmentioning

confidence: 99%

MKL-tree: an index structure for high-dimensional vector spaces

2006

View full text Add to dashboard Cite

In this work, a novel hierarchical data structure for high dimensional data indexing is proposed. MKL-tree is based on dimensionality reduction operated by means of the MKL transform, a multi-space generalization of the KL transform. A local dimensionality reduction is performed at each node of the tree, allowing more selective features to be extracted and thus increasing the discriminating power of the index. The mathematical foundation for nodes and leaves representation and for the techniques aimed to manage the structure is detailed. Moreover, the algorithms for bulk loading MKL-tree (i.e., for creating the tree given a large number of objects simultaneously), for updating and splitting nodes after the insertion of new objects and for performing similarity searches are described. Results are reported for the comparison of MKL-tree with other well-known access methods in terms of I/O and CPU costs and precision of the result in the execution of similarity queries.

show abstract

“…Cells which having more than specified numbers of points are called dense cells. The dense cells are connected to form clusters [13], some of which are: STING [16] and CLIQUE [17].…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Approach for Data Clustering using Expectation-Maximization and Parameter Adaptive Harmony Search Algorithm

2015

March 29-30, 2015 Singapore

View full text Add to dashboard Cite

This paper presents a novel hybrid data clustering algorithm based on parameter adaptive harmony search algorithm. The recently developed parameter adaptive harmony search algorithm (PAHS) is used to refine the cluster centers, which are further used in initializing Expectation-Maximization clustering algorithm. The optimal number of clusters are determined through four well-known cluster validity indices. The proposed algorithm is evaluated on three real life datasets and compared with the performance of K-Means, Fuzzy CMeans and HS initialize EM (HSEM). Experimental results reveal that the proposed approach provide better results in terms of precision, recall, weighted average, F-Measure and G-Measure.

show abstract

Automatic subspace clustering of high dimensional data for data mining applications

Cited by 1,569 publications

References 25 publications

Efficient approaches for summarizing subspace clusters into k representatives

Efficient approaches for summarizing subspace clusters into k representatives

MKL-tree: an index structure for high-dimensional vector spaces

A Hybrid Approach for Data Clustering using Expectation-Maximization and Parameter Adaptive Harmony Search Algorithm

Contact Info

Product

Resources

About