Mixed integer linear programming and heuristic methods for feature selection in clustering

被引:13
作者
Benati, Stefano [1 ]
Garcia, Sergio [2 ]
Puerto, Justo [3 ]
机构
[1] Univ Trento, Sch Int Studies, Trento, Italy
[2] Univ Edinburgh, Sch Math, Edinburgh, Midlothian, Scotland
[3] Univ Seville, IMUS, Seville, Spain
关键词
Integer linear programming; heuristics; q-yars; cluster analysis; p-median problem; VARIABLE SELECTION; LOCATION-PROBLEMS; MODEL; FORMULATION; ALGORITHM; CUT;
D O I
10.1080/01605682.2017.1398206
中图分类号
C93 [管理学];
学科分类号
12 ; 1201 ; 1202 ; 120202 ;
摘要
This paper studies the problem of selecting relevant features in clustering problems, out of a data-set in which many features are useless, or masking. The data-set comprises a set U of units, a set V of features, a set R of (tentative) cluster centres and distances d(ijk) for every i is an element of U, k is an element of R, j is an element of V. The feature selection problem consists of finding a subset of features Q subset of V such that the total sum of the distances from the units to the closest centre is minimised. This is a combinatorial optimisation problem that we show to be NP-complete, and we propose two mixed integer linear programming formulations to calculate the solution. Some computational experiments show that if clusters are well separated and the relevant features are easy to detect, then both formulations can solve problems with many integer variables. Conversely, if clusters overlap and relevant features are ambiguous, then even small problems are unsolved. To overcome this difficulty, we propose two heuristic methods to find that, most of the time, one of them, called q-vars, calculates the optimal solution quickly. Then, the q-vars heuristic is combined with the k-means algorithm to cluster some simulated data. We conclude that this approach outperforms other methods for clustering with variable selection that were proposed in the literature.
引用
收藏
页码:1379 / 1395
页数:17
相关论文
共 44 条
[31]   Feature selection for Support Vector Machines via Mixed Integer Linear Programming [J].
Maldonado, Sebastian ;
Perez, Juan ;
Weber, Richard ;
Labbe, Martine .
INFORMATION SCIENCES, 2014, 279 :163-175
[32]   A flexible model and efficient solution strategies for discrete location problems [J].
Marin, Alfredo ;
Nickel, Stefan ;
Puerto, Justo ;
Velten, Sebastian .
DISCRETE APPLIED MATHEMATICS, 2009, 157 (05) :1128-1145
[33]  
McLachlan G., 1997, The EM Algorithm and Extensions
[34]   The p-median problem:: A survey of metaheuristic approaches [J].
Mladenovic, Nenad ;
Brimberg, Jack ;
Hansen, Pierre ;
Moreno-Perez, Jose A. .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2007, 179 (03) :927-939
[35]  
Morlini I., 2013, CLASSIFICATION DATA, P71
[36]  
Pan W., 2007, J MACHINE LEARNING R, V8, P1154
[37]   A specialized branch & bound & cut for Single-Allocation Ordered Median Hub Location problems [J].
Puerto, J. ;
Ramos, A. B. ;
Rodriguez-Chia, A. M. .
DISCRETE APPLIED MATHEMATICS, 2013, 161 (16-17) :2624-2646
[38]   Generation of random clusters with specified degree of separation [J].
Qiu, Weiliang ;
Joe, Harry .
JOURNAL OF CLASSIFICATION, 2006, 23 (02) :315-334
[39]   Variable selection for model-based clustering [J].
Raftery, AE ;
Dean, N .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2006, 101 (473) :168-178
[40]   Selection of variables in cluster analysis: An empirical comparison of eight procedures [J].
Steinley, Douglas ;
Brusco, Michael J. .
PSYCHOMETRIKA, 2008, 73 (01) :125-144