Improvements to Platt's SMO Algorithm for SVM Classifier Design

By S. Sathiya Keerthi, Shirish K. Shevade, C. Bhattacharyya, K. Murthy. Neural Computation, 2001

1,974 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering, Health Information Management

View Paper: https://www.semanticscholar.org/paper/6d95b96d71669f3f4edfcc95cacd428b62b3fcde

How researchers see this paper

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

In Artificial Intelligence

In Control and Systems Engineering

In Health Information Management

Citations in the corpus

A Study on Sigmoid Kernels for SVM and the Training of non-PSD Kernels by SMO-type Methods (2005, methodology): “A simple and common way to select the two variables is through the following form of optimal conditions (Keerthi, Shevade, Bhattacharyya, and Murthy 2001; Chang and Lin 2001): α is a local optimum of (1.2) if and only if α is feasible and max t∈Iup(α,C) −yt∇F (α)t ≤ min t∈Ilow(α,C) −yt∇F (α)t,…”
Convergence of a Generalized SMO Algorithm for SVM Classifier Design (2002, methodology): “Let us now briefly discuss the two modified SMO algorithms given by Keerthi et al. (1999) .”
A smart antenna system for direction of arrival estimation based on a support vector regression (2005, methodology): “Moreover, fast iterative algorithms based on the use of SVMs and relatively simple to be implemented, have been developed (see, for example, [12], [13] and the references therein).”
Transition-Aware Human Activity Recognition Using Smartphones (2016, methodology): “where the bias term b is obtained with the method proposed in [42].”
Kernel Methods and Machine Learning (2014, background): “) For suggestions on further improving Platt’s SMO algorithm, see [122].”
Human Activity Recognition on Smartphones Using a Multiclass Hardware-Friendly Support Vector Machine (2012, methodology): “where b is the bias term and is obtained by using the method proposed in [15].”
Feature Selection With Harmony Search (2012, methodology): “Note that, for data set ozone, the classifiers VQNN and SVM obtained identical cross-validated accuracy of 93.69%.”
A dual coordinate descent method for large-scale linear SVM (2008, background): “However, subsequent developments (Joachims, 1998; Chang & Lin, 2001; Keerthi et al., 2001) all use gradient information to conduct the selection.”
Tikhonov, Ivanov and Morozov regularization for support vector machine learning (2016, methodology): “7, we show experimentally the advantages and disadvantages of using the wellknown T-SVM Sequential Minimal Optimization (SMO) solver (Platt 1998; Keerthi et al. 2001; Keerthi and Gilbert 2002; Fan et al. 2005) for addressing both I-SVM and M-SVM problems.”
Support vector machines: hype or hallelujah? (2000, background): “SMO and improved versions [33] have proven to be an effective approach for large problems.”
Multiple kernel learning, conic duality, and the SMO algorithm (2004, methodology): “…the MY-regularized SKM are exactly the same as for the SVM, but with a diﬀerent objective function (Platt, 1998; Keerthi et al., 2001): choosing the pair of indices can be done in a manner similar to that proposed for the SVM, by using the fast heuristics of Platt (1998) and Keerthi et al. (2001).”
Gaussian Processes for Ordinal Regression (2005, methodology): “7 The improved SMO algorithm (Keerthi et al., 2001) was adapted to implement the SVM approach (refer to Chu and Keerthi (2005) for detailed description and extensive discussion),8 and 5-fold cross validation was used to determine the optimal values of model parameters (the kernel parameter κ and the”
LIBSVM: A library for support vector machines (2011, neutral): “Some earlier works on decomposition methods for SVM include, for example, Osuna et al. (1997a); Joachims (1998); Platt (1998); Keerthi et al. (2001); Hsu and Lin (2002b).”
A Study on SMO-Type Decomposition Methods for Support Vector Machines (2006, positive): “This working set was first proposed in [14] and has been used in many software packages, for example, LIBSVM [3].”