A Pattern Search Method for Model Selection of Support Vector Regression

By Michinari Momma, Kristin P. Bennett. 2002

177 citations recorded by the source corpus.

Subfields: Analytical Chemistry, Computational Theory and Mathematics, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Analytical Chemistry

In Computational Theory and Mathematics

In Computer Vision and Pattern Recognition

Citations in the corpus

Benchmark and Survey of Automated Machine Learning Frameworks (2019, neutral): “Many ML models provide some sort of feature ranking that can be utilized, e.g., SVMs (Guyon et al., 2002; Rakotomamonjy, 2003), perceptrons (Mej́ıa-Lavalle et al., 2006) or random forests (Tuv et al., 2009).”
Model Selection: Beyond the Bayesian/Frequentist Divide (2010, background): “These include sequential search strategies such as coordinate ascent or descent (making small steps along coordinate axes) or pattern search (Momma and Bennett, 2002) (making local steps according to a certain pattern), which, by accepting only moves that improve the evaluation function, find the lo”
Analysis of the AutoML Challenge Series 2015-2018 (2019, background): “It explores the hyper-parameter space by steps of the same magnitude, and when no change in any parameter further decreases J2, the step size is halved and the process repeated until the steps are deemed sufficiently small (Momma and Bennett, 2002).”
A nested heuristic for parameter tuning in Support Vector Machines (2014, background): “An early reference on metaheuristics is [43], where a Pattern Search approach is introduced.”
Meta-learning for evolutionary parameter optimization of classifiers (2012, methodology): “For example, to optimize parameters of an SVM, Pattern Search (Kolda et al. 2003) was used by Momma and Bennett (2002) as well as by Eitrich and Lang (2006), and a gradient descent-based approach was used by Chapelle et al. (2002).”
Design of the 2015 ChaLearn AutoML challenge (2015, background): “It explores hyperparameter space by steps of the same magnitude, and when no change in any one parameter further decreases J2, the step size is halved and the process repeated until the steps are deemed sufficiently small [52].”
Particle Swarm Model Selection for Authorship Verification (2009, background): “Traditionally, a single learning algorithm is considered and the task is to optimize the model’s parameters such that its generalization performance is maximized [12].”
Supervised classification and mathematical optimization (2013, methodology): “These include pattern search methods, [178] or evolutionary algorithms, [63, 97, 152].”