Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Clustering in the presence of side information: a non-linear approach

Clustering in the presence of side information: a non-linear approach Constrained clustering is an important recent development in clustering literature. The goal of an algorithm in constrained clustering research is to improve the quality of clustering by making use of background knowledge. The purpose of this paper is to suggest a new perspective for constrained clustering, by finding an effective transformation of data into target space on the reference of background knowledge given in the form of pairwise must- and cannot-link constraints.Design/methodology/approachMost of existing methods in constrained clustering are limited to learn a distance metric or kernel matrix from the background knowledge while looking for transformation of data in target space. Unlike previous efforts, the author presents a non-linear method for constraint clustering, whose basic idea is to use different non-linear functions for each dimension in target space.FindingsThe outcome of the paper is a novel non-linear method for constrained clustering which uses different non-linear functions for each dimension in target space. The proposed method for a particular case is formulated and explained for quadratic functions. To reduce the number of optimization parameters, the proposed method is modified to relax the quadratic function and approximate it by a factorized version that is easier to solve. Experimental results on synthetic and real-world data demonstrate the efficacy of the proposed method.Originality/valueThis study proposes a new direction to the problem of constrained clustering by learning a non-linear transformation of data into target space without using kernel functions. This work will assist researchers to start development of new methods based on the proposed framework which will potentially provide them with new research topics. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Intelligent Computing and Cybernetics Emerald Publishing

Clustering in the presence of side information: a non-linear approach

Download PDF
23 pages

Loading next page...
 
/lp/emerald-publishing/clustering-in-the-presence-of-side-information-a-non-linear-approach-LwrwgzD7yZ
Publisher
Emerald Publishing
Copyright
© Emerald Publishing Limited
ISSN
1756-378X
DOI
10.1108/ijicc-04-2018-0046
Publisher site
See Article on Publisher Site

Abstract

Constrained clustering is an important recent development in clustering literature. The goal of an algorithm in constrained clustering research is to improve the quality of clustering by making use of background knowledge. The purpose of this paper is to suggest a new perspective for constrained clustering, by finding an effective transformation of data into target space on the reference of background knowledge given in the form of pairwise must- and cannot-link constraints.Design/methodology/approachMost of existing methods in constrained clustering are limited to learn a distance metric or kernel matrix from the background knowledge while looking for transformation of data in target space. Unlike previous efforts, the author presents a non-linear method for constraint clustering, whose basic idea is to use different non-linear functions for each dimension in target space.FindingsThe outcome of the paper is a novel non-linear method for constrained clustering which uses different non-linear functions for each dimension in target space. The proposed method for a particular case is formulated and explained for quadratic functions. To reduce the number of optimization parameters, the proposed method is modified to relax the quadratic function and approximate it by a factorized version that is easier to solve. Experimental results on synthetic and real-world data demonstrate the efficacy of the proposed method.Originality/valueThis study proposes a new direction to the problem of constrained clustering by learning a non-linear transformation of data into target space without using kernel functions. This work will assist researchers to start development of new methods based on the proposed framework which will potentially provide them with new research topics.

Journal

International Journal of Intelligent Computing and Cybernetics – Emerald Publishing

Published: May 15, 2019

Keywords: Cannot-link; Constrained clustering; Instance-level constraints; Must-link; Quadratic functions; Side information

References

  • Querying beneficial constraints before clustering using facility location analysis
    Abin, A.A.
  • Clustering with side information: further efforts to improve efficiency
    Abin, A.A.
  • Active constrained fuzzy clustering: a multiple kernels learning approach
    Abin, A.A.; Beigy, H.
  • Synchronization in complex networks
    Arenas, A.; Diaz-Guilera, A.; Kurths, J.; Moreno, Y.; Zhou, C.
  • Learning a Mahalanobis metric from equivalence constraints
    Bar-Hillel, A.; Hertz, T.; Shental, N.; Weinshall, D.
  • Active semi-supervision for pairwise constrained clustering
    Basu, S.; Banerjee, A.; Mooney, R.J.
  • Integrating constraints and metric learning in semi-supervised clustering
    Bilenko, M.; Basu, S.; Mooney, R.J.
  • Locally linear metric adaptation for semi-supervised clustering
    Chang, H.; yan Yeung, D.
  • A survey of clustering with instance level constraints
    Davidson, I.; Basu, S.
  • Evaluation of employee profiles using a hybrid clustering and optimization model: practical study
    Esmaeilzadeh, M.; Abdollahi, B.; Ganjali, A.; Hasanpoor, A.
  • Combined data mining techniques based patient data outlier detection for healthcare safety
    Gebremeskel, G.B.; Yi, C.; He, Z.; Haile, D.
  • Boosting margin based distance functions for clustering
    Hertz, T.; Bar-hillel, A.; Weinshall, D.
  • Learning non-parametric kernel matrices from pairwise constraints
    Hoi, S.C.H.; Jin, R.; Lyu, M.R.
  • Learning distance metrics with contextual constraints for image retrieval
    Hoi, S.C.H.; Liu, W.; Lyu, M.R.; Ma, W.-Y.
  • Metric and kernel learning using a linear transformation
    Jain, P.; Kulis, B.; Davis, J.V.; Dhillon, I.S.
  • Pairwise constraint propagation by semi-definite programming for semi-supervised classification
    Li, Z.; Liu, J.; Tang, X.
  • Constrained metric learning via distance gap maximization
    Liu, W.; Tian, X.; Tao, D.; Liu, J.
  • Semi-supervised model-based clustering with positive and negative constraints
    Melnykov, V.; Melnykov, I.; Michael, S.
  • Clustering with constrained similarity learning
    Okabe, M.; Yamada, S.
  • Clustering using boosted constrained k-means algorithm
    Okabe, M.; Yamada, S.
  • Constrained clustering with a complex cluster structure
    Smieja, M.; Wiercioch, M.
  • Non-linear metric learning using pairwise similarity and dissimilarity constraints and the geometrical structure of data
    Soleymani Baghshah, M.; Bagheri Shouraki, S.
  • Scalable semi-supervised clustering by spectral kernel learning
    Soleymani Baghshah, M.; Afsari, F.; Bagheri Shouraki, S.; Eslami, E.
  • Cluster ensembles – a knowledge reuse framework for combining multiple partitions
    Strehl, A.; Ghosh, J.
  • Distance metric learning with kernels
    Tsang, I.W.; Kwok, J.T.
  • Kernel relevant component analysis for distance metric learning
    Tsang, I.W.; Kwok, J.T.
  • Value, cost, and sharing: open issues in constrained clustering
    Wagstaff, K.
  • Clustering with instance-level constraints
    Wagstaff, K.; Cardie, C.
  • Constrained k-means clustering with background knowledge
    Wagstaff, K.; Cardie, C.; Rogers, S.; SchrOdl, S.
  • Active spectral clustering
    Wang, X.; Davidson, I.
  • Exploring topics related to data mining on Wikipedia
    Wang, Y.; Zhang, J.
  • Learning a Mahalanobis distance metric for data clustering and classification
    Xiang, S.; Nie, F.; Zhang, C.
  • Distance metric learning with application to clustering with side-information
    Xing, E.P.; Ng, A.Y.; Jordan, M.I.; Russell, S.J.
  • Adaptive distance metric learning for clustering
    Ye, J.; Zhao, Z.; Liu, H.
  • Semi-supervised clustering with metric learning: an adaptive kernel method
    Yin, X.; Chen, S.; Hu, E.; Zhang, D.
  • Improved possibilistic c-means clustering algorithms
    Zhang, J.-S.; Leung, Y.-W.
  • Semi-supervised spectral clustering algorithm based on Bayesian decision
    Zhou, Y.; Wang, Y.; Chen, D.; Liu, B.
  • Semi-supervised learning literature survey
    Zhu, X.