A meta-learning recommender system for hyperparameter tuning: Predicting when tuning improves SVM classifiers

Information Sciences - Tập 501 - Trang 193-221 - 2019
Rafael Gomes Mantovani1,2, André Luis Debiaso Rossi3, Edesio Alcobaça2, Joaquin Vanschoren4, André C. P. L. F. de Carvalho2
1Federal Technology University, Campus of Apucarana, PR, Brazil
2Institute of Mathematics and Computer Sciences, University of São Paulo, São Carlos, SP, Brazil
3Universidade Estadual Paulista, Campus de Itapeva, São Paulo, Brazil
4Eindhoven University of Technology, Eindhoven, The Netherlands

Tóm tắt

Từ khóa


Tài liệu tham khảo

Ali, 2006, A meta-learning approach to automatic kernel selection for support vector machines, Neurocomputing, 70, 173, 10.1016/j.neucom.2006.03.004

Bensusan, 2000, A higher-order approach to meta-learning, 109

Bergstra, 2012, Random search for hyper-parameter optimization, J. Mach. Learn. Res., 13, 281

Bischl, 2016, Mlr: machine learning in R, J. Mach. Learn. Res., 17, 1

Brazdil, 2009

Brazdil, 1994, Analysis of results, 175

Breiman, 2001, Random forests, Mach. Learn., 45, 5, 10.1023/A:1010933404324

Brodersen, 2010, The balanced accuracy and its posterior distribution, 3121

Casalicchio, 2017, Openml: an R package to connect to the machine learning platform openml, Comput. Stat., 1

Chang, 2011, LIBSVM: a library for support vector machines, ACM Trans. Intell. Syst.Technol., 2, 27:1, 10.1145/1961189.1961199

Chawla, 2002, SMOTE: synthetic minority over-sampling technique, J. Artif. Int. Res., 16, 321

Demšar, 2006, Statistical comparisons of classifiers over multiple data sets, J. Mach. Learn. Res., 7, 1

Eggensperger, 2018, Efficient benchmarking of algorithm configurators via model-based surrogates, Mach. Learn., 107, 15, 10.1007/s10994-017-5683-z

Feurer, 2015, Efficient and robust automated machine learning, 2944

Garcia, 2016, Noise detection in the meta-learning level, Neurocomputing, 176, 14, 10.1016/j.neucom.2014.12.100

Gomes, 2012, Combining meta-learning and search techniques to select parameters for support vector machines, Neurocomputing, 75, 3, 10.1016/j.neucom.2011.07.005

Ho, 2002, Complexity measures of supervised classification problems, Pattern Anal. Mach. Intell. IEEE Trans., 24, 289, 10.1109/34.990132

Horn, 2016, A comparative study on large scale kernelized support vector machines, Adv. Data Anal. Classif., 1

Hsu, 2007

Hutter, 2009, Paramils: an automatic algorithm configuration framework, J. Artif. Intell. Res., 267, 10.1613/jair.2861

Kotthoff, 2016, Auto-weka 2.0: automatic model selection and hyperparameter optimization in weka, J. Mach. Learn. Res., 17, 1

Krstajic, 2014, Cross-validation pitfalls when selecting and assessing regression and classification models., J.Cheminform., 6, 10+, 10.1186/1758-2946-6-10

Leite, 2012, Selecting classification algorithms with active testing, 117

López-Ibáñez, 2016, The irace package: iterated racing for automatic algorithm configuration, Oper. Res. Perspect., 3, 43, 10.1016/j.orp.2016.09.002

Lorena, 2018, Data complexity meta-features for regression problems, Mach. Learn., 107, 209, 10.1007/s10994-017-5681-1

Mantovani, 2016, Hyper-parameter tuning of a decision tree induction algorithm, 37

Mantovani, 2015, Effectiveness of random search in SVM hyper-parameter tuning, 1

Mantovani, 2015, To tune or not to tune: Recommending when to adjust SVM hyper-parameters via meta-learning, 1

Mantovani, 2015, Meta-learning recommendation of default hyper-parameter values for svms in classification tasks, 1455, 80

Miranda, 2013, Active testing for SVM parameter selection, 1

Miranda, 2014, Fine-tuning of support vector machine parameters using racing algorithms, 325

Miranda, 2012, An experimental study of the combination of meta-learning with particle swarm algorithms for svm parameter selection, Lect. Notes Comput. Sci., 7335, 562, 10.1007/978-3-642-31137-6_43

Morais, 2013, Complex network measures for data set characterization, 12

L.C. Padierna, M. Carpio, A. Rojas, H. Puga, R. Baltazar, H. Fraire, Hyper-Parameter Tuning for Support Vector Machines by Estimation of Distribution Algorithms, Springer International Publishing, Cham, pp. 787–800.

Pfahringer, 2000, Meta-learning by landmarking various learning algorithms, 743

Priya, 2012, Using genetic algorithms to improve prediction of execution times of ML tasks, 7208 LNAI, 196

Quinlan, 1993

Reif, 2011, Prediction of classifier training time including parameter optimization, 260

Reif, 2012, Meta-learning for evolutionary parameter optimization of classifiers, Mach. Learn., 87, 357, 10.1007/s10994-012-5286-7

Reif, 2014, Automatic classifier selection for non-experts, Pattern Anal. Appl., 17, 83, 10.1007/s10044-012-0280-z

Ridd, 2014, Using metalearning to predict when parameter optimization is likely to improve classification accuracy, 18

Sanders, 2017, Informing the use of hyperparameter optimization through metalearning, 1051

Snoek, 2012, Practical Bayesian optimization of machine learning algorithms, 2951

Soares, 2006, Selecting parameters of svm using meta-learning and kernel matrix-based meta-features, 564

Soares, 2004, A meta-learning method to select the kernel width in support vector regression, Mach. Learn., 54, 195, 10.1023/B:MACH.0000015879.28004.9b

Thornton, 2013, Auto-WEKA: combined selection and hyperparameter optimization of classification algorithms, 847

Vanschoren, 2014, Openml: networked science in machine learning, SIGKDD Explor. Newsl., 15, 49, 10.1145/2641190.2641198

Vapnik, 1995

Vilalta, 2004, Using meta-learning to support data mining, Int. J. Comput. Sci.Appl., 1, 31

Wistuba, 2018, Scalable gaussian process-based transfer surrogates for hyperparameter optimization, Mach. Learn., 107, 43, 10.1007/s10994-017-5684-y