# American Institute of Mathematical Sciences

doi: 10.3934/jimo.2020160

## The approximation algorithm based on seeding method for functional $k$-means problem†

 1 School of Mathematics and Statistics, Shandong Normal University, Jinan 250014, China 2 Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, 1068 Xueyuan Avenue, Shenzhen University Town, Shenzhen 518055, China 3 Department of Operations Research and Information Engineering, Beijing University of Technology, Beijing 100124, China 4 School of Computer Science and Technology, Shandong Jianzhu University, Jinan 250101, China

* Corresponding author: Dongmei Zhang

A preliminary version of this article appeared in Proceedings of COCOON 2019, pp. 387-396.

Received  March 2020 Revised  July 2020 Published  November 2020

Different from the classical $k$-means problem, the functional $k$-means problem involves a kind of dynamic data, which is generated by continuous processes. In this paper, we mainly design an $O(\ln\; k)$-approximation algorithm based on the seeding method for functional $k$-means problem. Moreover, the numerical experiment presented shows that this algorithm is more efficient than the functional $k$-means clustering algorithm.

Citation: Min Li, Yishui Wang, Dachuan Xu, Dongmei Zhang. The approximation algorithm based on seeding method for functional $k$-means problem. Journal of Industrial & Management Optimization, doi: 10.3934/jimo.2020160
##### References:
 [1] C. Abraham, P. A. Cornillon, E. Matzner-Løber and N. Molinari, Unsupervised curve clustering using B-splines, Scandinavian Journal of Statistics, 30 (2003), 581-595.  doi: 10.1111/1467-9469.00350.  Google Scholar [2] S. Ahmadian, A. Norouzi-Fard, O. Svensson and J. Ward, Better guarantees for $k$-means and Euclidean $k$-median by primal-dual algorithms, SIAM Journal on Computing, (2019), FOCS17-97–FOCS17-156. doi: 10.1137/18M1171321.  Google Scholar [3] D. Aloise, A. Deshpande, P. Hansen and P. Popat, NP-hardness of Euclidean sum-of-squares clustering, Machine Learning, 75 (2009), 245-248.  doi: 10.1007/s10994-009-5103-0.  Google Scholar [4] D. Arthur and S. Vassilvitskii, $K$-means++: The advantages of careful seeding, Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms, SIAM, (2007), 1027–1035.  Google Scholar [5] M. Boullé, Functional data clustering via piecewise constant nonparametric density estimation, Pattern Recognition, 45 (2012), 4389-4401.   Google Scholar [6] C. Bouveyron and C. Brunet-Saumard, Model-based clustering of high-dimensional data: A review, Computational Statistics & Data Analysis, 71 (2014), 52-78.  doi: 10.1016/j.csda.2012.12.008.  Google Scholar [7] R. Gamasaee and M. Zarandi, A new dirichlet process for mining dynamic patterns in functional data, Information Sciences, 405 (2017), 55-80.  doi: 10.1016/j.ins.2017.04.008.  Google Scholar [8] S. Har-Peled and B. Sadri, How fast is the $k$-means method?, Algorithmica, 41 (2005), 185-202.  doi: 10.1007/s00453-004-1127-9.  Google Scholar [9] J. Jacques and C. Preda, Functional data clustering: A survey, Advances in Data Analysis and Classification, 8 (2014), 231-255.  doi: 10.1007/s11634-013-0158-y.  Google Scholar [10] S. Ji, D. Xu, L. Guo, M. Li and D. Zhang, The seeding algorithm for spherical $k$-means clustering with penalties, Journal of Combinatorial Optimization, 2020. doi: 10.1007/s10878-020-00569-1.  Google Scholar [11] M. Kayano, K. Dozono and S. Konishi, Functional cluster analysis via orthonormalized Gaussian basis expansions and its application, Journal of Classification, 27 (2010), 211-230.  doi: 10.1007/s00357-010-9054-8.  Google Scholar [12] M. Li, The bi-criteria seeding algorithms for two variants of $k$-means problem, Journal of Combinatorial Optimization, 2020. doi: 10.1007/s10878-020-00537-9.  Google Scholar [13] M. Li, D. Xu, J. Yue, D. Zhang and P. Zhang, The seeding algorithm for $k$-means problem with penalties, Journal of Combinatorial Optimization, 39 (2020), 15-32.  doi: 10.1007/s10878-019-00450-w.  Google Scholar [14] S. Lloyd, Least squares quantization in PCM, IEEE Transactions on Information Theory, 28 (1982), 129-137.  doi: 10.1109/TIT.1982.1056489.  Google Scholar [15] Y. Meng, J. Liang, F. Cao and Y. He, A new distance with derivative information for functional $k$-means clustering algorithm, Information Sciences, 463/464 (2018), 166-185.  doi: 10.1016/j.ins.2018.06.035.  Google Scholar [16] R. Ostrovsky, Y. Rabani, L. Schulman and C. Swamy, The effectiveness of Lloyd-type methods for the $k$-means problem, Journal of the ACM, 59 (2012), 28: 1–28: 22. doi: 10.1145/2395116.2395117.  Google Scholar [17] G. Ozturk and M. Ciftci, Clustering based polyhedral conic functions algorithm in classification, Journal of Industrial & Management Optimization, 11 (2015), 921-932.  doi: 10.3934/jimo.2015.11.921.  Google Scholar [18] J. Park and J. Ahn, Clustering multivariate functional data with phase variation, Biometrics, 73 (2017), 324-333.  doi: 10.1111/biom.12546.  Google Scholar [19] J. Peng and H. G. Müller, Distance-based clustering of sparsely observed stochastic processes, with applications to online auctions, The Annals of Applied Statistics, 2 (2008), 1056-1077.  doi: 10.1214/08-AOAS172.  Google Scholar [20] C. Preda, G. Saporta and C. Lévéder, PLS classification of functional data, Computational Statistics, 22 (2007), 223-235.  doi: 10.1007/s00180-007-0041-4.  Google Scholar [21] T. Tarpey and K. K. Kinateder, Clustering functional data, Journal of Classification, 20 (2003), 93-114.  doi: 10.1007/s00357-003-0007-3.  Google Scholar [22] D. Wei, A constant-factor bi-criteria approximation guarantee for $k$-means++, Proceedings of the Thirtieth International Conference on Neural Information Processing Systems, (2016), 604–612. Google Scholar [23] X. Wu, V. Kumar, J. Quinlan, J. Ross Ghosh, Q. Yang, H. Motoda, G. J. McLachlan, A. Ng, B. Liu, P.S. Yu, Z. H. Zhou, M. Steinbach, D. J. Hand and D. Steinberg, Top 10 algorithms in data mining, Knowledge and Information Systems, 14 (2008), 1-37.  doi: 10.1007/s10115-007-0114-2.  Google Scholar

show all references

##### References:
 [1] C. Abraham, P. A. Cornillon, E. Matzner-Løber and N. Molinari, Unsupervised curve clustering using B-splines, Scandinavian Journal of Statistics, 30 (2003), 581-595.  doi: 10.1111/1467-9469.00350.  Google Scholar [2] S. Ahmadian, A. Norouzi-Fard, O. Svensson and J. Ward, Better guarantees for $k$-means and Euclidean $k$-median by primal-dual algorithms, SIAM Journal on Computing, (2019), FOCS17-97–FOCS17-156. doi: 10.1137/18M1171321.  Google Scholar [3] D. Aloise, A. Deshpande, P. Hansen and P. Popat, NP-hardness of Euclidean sum-of-squares clustering, Machine Learning, 75 (2009), 245-248.  doi: 10.1007/s10994-009-5103-0.  Google Scholar [4] D. Arthur and S. Vassilvitskii, $K$-means++: The advantages of careful seeding, Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms, SIAM, (2007), 1027–1035.  Google Scholar [5] M. Boullé, Functional data clustering via piecewise constant nonparametric density estimation, Pattern Recognition, 45 (2012), 4389-4401.   Google Scholar [6] C. Bouveyron and C. Brunet-Saumard, Model-based clustering of high-dimensional data: A review, Computational Statistics & Data Analysis, 71 (2014), 52-78.  doi: 10.1016/j.csda.2012.12.008.  Google Scholar [7] R. Gamasaee and M. Zarandi, A new dirichlet process for mining dynamic patterns in functional data, Information Sciences, 405 (2017), 55-80.  doi: 10.1016/j.ins.2017.04.008.  Google Scholar [8] S. Har-Peled and B. Sadri, How fast is the $k$-means method?, Algorithmica, 41 (2005), 185-202.  doi: 10.1007/s00453-004-1127-9.  Google Scholar [9] J. Jacques and C. Preda, Functional data clustering: A survey, Advances in Data Analysis and Classification, 8 (2014), 231-255.  doi: 10.1007/s11634-013-0158-y.  Google Scholar [10] S. Ji, D. Xu, L. Guo, M. Li and D. Zhang, The seeding algorithm for spherical $k$-means clustering with penalties, Journal of Combinatorial Optimization, 2020. doi: 10.1007/s10878-020-00569-1.  Google Scholar [11] M. Kayano, K. Dozono and S. Konishi, Functional cluster analysis via orthonormalized Gaussian basis expansions and its application, Journal of Classification, 27 (2010), 211-230.  doi: 10.1007/s00357-010-9054-8.  Google Scholar [12] M. Li, The bi-criteria seeding algorithms for two variants of $k$-means problem, Journal of Combinatorial Optimization, 2020. doi: 10.1007/s10878-020-00537-9.  Google Scholar [13] M. Li, D. Xu, J. Yue, D. Zhang and P. Zhang, The seeding algorithm for $k$-means problem with penalties, Journal of Combinatorial Optimization, 39 (2020), 15-32.  doi: 10.1007/s10878-019-00450-w.  Google Scholar [14] S. Lloyd, Least squares quantization in PCM, IEEE Transactions on Information Theory, 28 (1982), 129-137.  doi: 10.1109/TIT.1982.1056489.  Google Scholar [15] Y. Meng, J. Liang, F. Cao and Y. He, A new distance with derivative information for functional $k$-means clustering algorithm, Information Sciences, 463/464 (2018), 166-185.  doi: 10.1016/j.ins.2018.06.035.  Google Scholar [16] R. Ostrovsky, Y. Rabani, L. Schulman and C. Swamy, The effectiveness of Lloyd-type methods for the $k$-means problem, Journal of the ACM, 59 (2012), 28: 1–28: 22. doi: 10.1145/2395116.2395117.  Google Scholar [17] G. Ozturk and M. Ciftci, Clustering based polyhedral conic functions algorithm in classification, Journal of Industrial & Management Optimization, 11 (2015), 921-932.  doi: 10.3934/jimo.2015.11.921.  Google Scholar [18] J. Park and J. Ahn, Clustering multivariate functional data with phase variation, Biometrics, 73 (2017), 324-333.  doi: 10.1111/biom.12546.  Google Scholar [19] J. Peng and H. G. Müller, Distance-based clustering of sparsely observed stochastic processes, with applications to online auctions, The Annals of Applied Statistics, 2 (2008), 1056-1077.  doi: 10.1214/08-AOAS172.  Google Scholar [20] C. Preda, G. Saporta and C. Lévéder, PLS classification of functional data, Computational Statistics, 22 (2007), 223-235.  doi: 10.1007/s00180-007-0041-4.  Google Scholar [21] T. Tarpey and K. K. Kinateder, Clustering functional data, Journal of Classification, 20 (2003), 93-114.  doi: 10.1007/s00357-003-0007-3.  Google Scholar [22] D. Wei, A constant-factor bi-criteria approximation guarantee for $k$-means++, Proceedings of the Thirtieth International Conference on Neural Information Processing Systems, (2016), 604–612. Google Scholar [23] X. Wu, V. Kumar, J. Quinlan, J. Ross Ghosh, Q. Yang, H. Motoda, G. J. McLachlan, A. Ng, B. Liu, P.S. Yu, Z. H. Zhou, M. Steinbach, D. J. Hand and D. Steinberg, Top 10 algorithms in data mining, Knowledge and Information Systems, 14 (2008), 1-37.  doi: 10.1007/s10115-007-0114-2.  Google Scholar
Notation Index
 Notations Meaning of symbols $x_i(t)$ or $x_i^j(t)$ functional curve $X(t)$ or $X^i(t)$ or $C^i(t)$ $d$-dimensional functional sample with functional curves as its components $\mathfrak{F}^d(t)$ set of $d$-dimensional functional samples $\Gamma(t)$ or $\Delta(t)$ or $C(t)$ subset of $\mathfrak{F}^d(t)$ $\mu(\Gamma(t))$ center of mass of functional samples of $\Gamma(t)$ $d(X^i(t), X^j(t))$ distance between the functional samples $X^i(t)$ and $X^j(t)$ $d(X^i(t), \Gamma(t))$ distance from the functional sample $X^i(t)$ to the subset of functional samples $\Gamma(t)$ $X(t)_{\Gamma(t)}$ the closest functional sample in $\Gamma(t)$ to the functional sample $X(t)$ $\Phi(\Gamma(t), C(t))$ potential function of $\Gamma(t)$ over $C(t)$ $\Gamma_{C(t)}^i(t)$ the $i$-th cluster of $\Gamma(t)$ with respect to $C(t)$
 Notations Meaning of symbols $x_i(t)$ or $x_i^j(t)$ functional curve $X(t)$ or $X^i(t)$ or $C^i(t)$ $d$-dimensional functional sample with functional curves as its components $\mathfrak{F}^d(t)$ set of $d$-dimensional functional samples $\Gamma(t)$ or $\Delta(t)$ or $C(t)$ subset of $\mathfrak{F}^d(t)$ $\mu(\Gamma(t))$ center of mass of functional samples of $\Gamma(t)$ $d(X^i(t), X^j(t))$ distance between the functional samples $X^i(t)$ and $X^j(t)$ $d(X^i(t), \Gamma(t))$ distance from the functional sample $X^i(t)$ to the subset of functional samples $\Gamma(t)$ $X(t)_{\Gamma(t)}$ the closest functional sample in $\Gamma(t)$ to the functional sample $X(t)$ $\Phi(\Gamma(t), C(t))$ potential function of $\Gamma(t)$ over $C(t)$ $\Gamma_{C(t)}^i(t)$ the $i$-th cluster of $\Gamma(t)$ with respect to $C(t)$
Comparison of Algorithm 1 and the functional $k$-means algorithm in [15]
 Data Set Method ARI DBI Initial Cost Returned Cost Time (s) Simudata SeedAlg 0.7919 0.8652 3338 1689 58 FuncAlg 0.7975 0.8656 5465 1689 61 Sdata SeedAlg 0.6651 1.1385 1667 712 192 FuncAlg 0.6561 1.1384 2485 720 217
 Data Set Method ARI DBI Initial Cost Returned Cost Time (s) Simudata SeedAlg 0.7919 0.8652 3338 1689 58 FuncAlg 0.7975 0.8656 5465 1689 61 Sdata SeedAlg 0.6651 1.1385 1667 712 192 FuncAlg 0.6561 1.1384 2485 720 217
 [1] J. Frédéric Bonnans, Justina Gianatti, Francisco J. Silva. On the convergence of the Sakawa-Shindo algorithm in stochastic control. Mathematical Control & Related Fields, 2016, 6 (3) : 391-406. doi: 10.3934/mcrf.2016008 [2] Demetres D. Kouvatsos, Jumma S. Alanazi, Kevin Smith. A unified ME algorithm for arbitrary open QNMs with mixed blocking mechanisms. Numerical Algebra, Control & Optimization, 2011, 1 (4) : 781-816. doi: 10.3934/naco.2011.1.781 [3] Ralf Hielscher, Michael Quellmalz. Reconstructing a function on the sphere from its means along vertical slices. Inverse Problems & Imaging, 2016, 10 (3) : 711-739. doi: 10.3934/ipi.2016018 [4] Davide La Torre, Simone Marsiglio, Franklin Mendivil, Fabio Privileggi. Public debt dynamics under ambiguity by means of iterated function systems on density functions. Discrete & Continuous Dynamical Systems - B, 2021  doi: 10.3934/dcdsb.2021070 [5] Ka Luen Cheung, Man Chun Leung. Asymptotic behavior of positive solutions of the equation $\Delta u + K u^{\frac{n+2}{n-2}} = 0$ in $IR^n$ and positive scalar curvature. Conference Publications, 2001, 2001 (Special) : 109-120. doi: 10.3934/proc.2001.2001.109 [6] Wen-Bin Yang, Yan-Ling Li, Jianhua Wu, Hai-Xia Li. Dynamics of a food chain model with ratio-dependent and modified Leslie-Gower functional responses. Discrete & Continuous Dynamical Systems - B, 2015, 20 (7) : 2269-2290. doi: 10.3934/dcdsb.2015.20.2269 [7] Zhihua Zhang, Naoki Saito. PHLST with adaptive tiling and its application to antarctic remote sensing image approximation. Inverse Problems & Imaging, 2014, 8 (1) : 321-337. doi: 10.3934/ipi.2014.8.321 [8] Luke Finlay, Vladimir Gaitsgory, Ivan Lebedev. Linear programming solutions of periodic optimization problems: approximation of the optimal control. Journal of Industrial & Management Optimization, 2007, 3 (2) : 399-413. doi: 10.3934/jimo.2007.3.399 [9] Xianming Liu, Guangyue Han. A Wong-Zakai approximation of stochastic differential equations driven by a general semimartingale. Discrete & Continuous Dynamical Systems - B, 2021, 26 (5) : 2499-2508. doi: 10.3934/dcdsb.2020192

2019 Impact Factor: 1.366

## Tools

Article outline

Figures and Tables