In-Sample and Out-of-Sample Model Selection and Error Estimation for Support Vector Machines

2012 ◽  
Vol 23 (9) ◽  
pp. 1390-1406 ◽  
Author(s):  
D. Anguita ◽  
A. Ghio ◽  
L. Oneto ◽  
S. Ridella
Keyword(s):  
Support Vector Machines ◽  
Model Selection ◽  
Error Estimation ◽  
Support Vector ◽  
Vector Machines ◽  
Out Of Sample

scholarly journals Asymptotic Behaviors of Support Vector Machines with Gaussian Kernel

2003 ◽  
Vol 15 (7) ◽  
pp. 1667-1689 ◽  
Author(s):  
S. Sathiya Keerthi ◽  
Chih-Jen Lin
Keyword(s):  
Support Vector Machines ◽  
Model Selection ◽  
Heuristic Method ◽  
Gaussian Kernel ◽  
Support Vector ◽  
Svm Classifier ◽  
Vector Machines ◽  
Rbf Kernel ◽  
Linear Svm ◽  
Generalization Errors

Support vector machines (SVMs) with the gaussian (RBF) kernel have been popular for practical use. Model selection in this class of SVMs involves two hyper parameters: the penalty parameter C and the kernel width σ. This letter analyzes the behavior of the SVM classifier when these hyper parameters take very small or very large values. Our results help in understanding the hyperparameter space that leads to an efficient heuristic method of searching for hyperparameter values with small generalization errors. The analysis also indicates that if complete model selection using the gaussian kernel has been conducted, there is no need to consider linear SVM.

scholarly journals Decomposition Methods for Linear Support Vector Machines

2004 ◽  
Vol 16 (8) ◽  
pp. 1689-1704 ◽  
Author(s):  
Wei-Chun Kao ◽  
Kai-Min Chung ◽  
Chia-Liang Sun ◽  
Chih-Jen Lin
Keyword(s):  
Support Vector Machines ◽  
Model Selection ◽  
Decomposition Methods ◽  
Dual Solutions ◽  
Support Vector ◽  
Vector Machines ◽  
Direct Use

In this letter, we show that decomposition methods with alpha seeding are extremely useful for solving a sequence of linear support vector machines (SVMs) with more data than attributes. This strategy is motivated by Keerthi and Lin (2003), who proved that for an SVM with data not linearly separable, after C is large enough, the dual solutions have the same free and bounded components. We explain why a direct use of decomposition methods for linear SVMs is sometimes very slow and then analyze why alpha seeding is much more effective for linear than nonlinear SVMs. We also conduct comparisons with other methods that are efficient for linear SVMs and demonstrate the effectiveness of alpha seeding techniques in model selection.

A comprehensive review on the variants of support vector machines

2019 ◽  
Vol 33 (25) ◽  
pp. 1950303 ◽  
Author(s):  
Bagesh Kumar ◽  
O. P. Vyas ◽  
Ranjana Vyas
Keyword(s):  
Machine Learning ◽  
Support Vector Machines ◽  
Model Selection ◽  
Support Vector ◽  
Automated Extraction ◽  
Future Directions ◽  
Comprehensive Review ◽  
Vector Machines ◽  
Unsupervised Approaches ◽  
Inputs And Outputs

Machine learning (ML) represents the automated extraction of models (or patterns) from data. All ML techniques start with data. These data describe the desired relationship between the ML model inputs and outputs, the latter of which may be implicit for unsupervised approaches. Equivalently, these data encode the requirements we wish to be embodied in our ML model. Thereafter, the model selection comes in action, to select an efficient ML model. In this paper, we have focused on various ML models which are the extensions of the well-known ML model, i.e. Support vector machines (SVMs). The main objective of this paper is to compare the existing ML models with the variants of SVM. Limitations of the existing techniques including the variants of SVM are then drawn. Finally, future directions are presented.

Sign in / Sign up

Export Citation Format

Share Document