An empirical approach for probing the definiteness of kernels

被引:0
作者
Zaefferer, Martin [1 ]
Bartz-Beielstein, Thomas [1 ]
Rudolph, Guenter [2 ]
机构
[1] TH Koln Univ Appl Sci, Fac Comp Sci & Engn Sci, Steinmullerallee 1, D-51643 Gummersbach, Germany
[2] TU Dortmund Univ, Dept Comp Sci, Otto Hahn Str 14, D-44227 Dortmund, Germany
关键词
Definiteness; Kernel; Distance; Sampling; Optimization; Evolutionary algorithm; SYMMETRIC-MATRICES; DISTANCE MEASURES; PERMUTATIONS; OPTIMIZATION; ALGORITHM; SEARCH;
D O I
10.1007/s00500-018-3648-1
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Models like support vector machines or Gaussian process regression often require positive semi-definite kernels. These kernels may be based on distance functions. While definiteness is proven for common distances and kernels, a proof for a new kernel may require too much time and effort for users who simply aim at practical usage. Furthermore, designing definite distances or kernels may be equally intricate. Finally, models can be enabled to use indefinite kernels. This may deteriorate the accuracy or computational cost of the model. Hence, an efficient method to determine definiteness is required. We propose an empirical approach. We show that sampling as well as optimization with an evolutionary algorithm may be employed to determine definiteness. We provide a proof of concept with 16 different distance measures for permutations. Our approach allows to disprove definiteness if a respective counterexample is found. It can also provide an estimate of how likely it is to obtain indefinite kernel matrices. This provides a simple, efficient tool to decide whether additional effort should be spent on designing/selecting a more suitable kernel or algorithm.
引用
收藏
页码:10939 / 10952
页数:14
相关论文
共 55 条
[1]  
[Anonymous], 1998, STAT LEARNING THEORY
[2]  
[Anonymous], 1984, GRAD TEXTS MATH
[3]  
[Anonymous], UCSCCRL9910 DEP COMP
[4]  
[Anonymous], 2004, P INT C MACH LEARN
[5]  
[Anonymous], 2012, MACHINE LEARNING
[6]  
Bader DA, 2004, GENOME REARRANGEMENT
[7]   Model-based methods for continuous and discrete global optimization [J].
Bartz-Beielstein, Thomas ;
Zaefferer, Martin .
APPLIED SOFT COMPUTING, 2017, 55 :154-167
[8]   SMS-EMOA: Multiobjective selection based on dominated hypervolume [J].
Beume, Nicola ;
Naujoks, Boris ;
Emmerich, Michael .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2007, 181 (03) :1653-1669
[9]  
Boytsov L., 2011, J. Exp. Algorithmics (JEA), V16, p1.1, DOI [10.1145/1963190.1963191, DOI 10.1145/1963190.1963191]
[10]   A tutorial on Support Vector Machines for pattern recognition [J].
Burges, CJC .
DATA MINING AND KNOWLEDGE DISCOVERY, 1998, 2 (02) :121-167