Spectral band selection is a fundamental problem in hyperspectral classification. This paper addresses the problem of band selection for hyperspectral remote sensing image and SVM parameter optimization. First, we present a thorough experime- ntal study to show the superiority of the generalization capability of the support vector machine (SVM) approach in the hyperspec- tral classification of remote sensing image. Second, we propose an evolutionary classification system based on particle swarm optimization (PSO) to improve the generalization performance of the SVM classifier. For this purpose, we have optimized the SVM classifier design by searching for the best value of the parameters that tune its discriminant function, and upstream by looking for the best subset of features that feed the classifier. The experiments are conducted on the basis of AVIRIS 92AV3C dataset. The obtained results clearly confirm the superiority of the SVM approach as compared to traditional classifiers, and suggest that further substantial improvements in terms of classification accuracy can be achieved by the proposed PSO- SVM classification system. important for a specific task. Some of them may be redundant or even irrelevant. Better performance may be achieved by discarding some features. In other circumstances, the dimensionality of input space may be decreased to save some computation effort, although this may slightly lower classification accuracy. Therefore, the classification process must be fast and accurate, using the smallest number of features. This objective can be achieved using feature selection. Feature selection strategies are often implied to explore the effect of irrelevant attributes on the performance of classifier systems. This study attempts to increase the classification accuracy rate by employing an approach based on particle swarm optimization (PSO) in SVM. This novel approach is termed PSO-SVM. The developed PSO-SVM approach not only tunes the parameter values of SVM, but also identifies a subset of features for specific problems, maximizing the classification accuracy rate of SVM. This makes the optimal separating hyper-plane obtainable in both linear and non-linear classification problems. In particular, they are organized so as to test the sensitivity of the SVM classifier and that of three reference classifiers used for comparison, i.e., SVM-Linear classifier, the k-nearest neighbor (K-nn) classifier and the radial basis function neural network (RBF-NN) classifier, with respect to the curse of dimensionality and the number of available training data
[1]
Riccardo Poli,et al.
Particle swarm optimization
,
1995,
Swarm Intelligence.
[2]
Bernhard Schölkopf,et al.
Feature selection for support vector machines by means of genetic algorithm
,
2003,
Proceedings. 15th IEEE International Conference on Tools with Artificial Intelligence.
[3]
Vladimir N. Vapnik,et al.
The Nature of Statistical Learning Theory
,
2000,
Statistics for Engineering and Information Science.
[4]
Cheng-Lung Huang,et al.
A GA-based feature selection and parameters optimizationfor support vector machines
,
2006,
Expert Syst. Appl..
[5]
J. Edward Jackson,et al.
A User's Guide to Principal Components.
,
1991
.
[6]
J. E. Jackson.
A User's Guide to Principal Components
,
1991
.
[7]
Chih-Jen Lin,et al.
LIBSVM: A library for support vector machines
,
2011,
TIST.
[8]
Vladimir Cherkassky,et al.
The Nature Of Statistical Learning Theory
,
1997,
IEEE Trans. Neural Networks.
[9]
James Kennedy,et al.
Particle swarm optimization
,
1995,
Proceedings of ICNN'95 - International Conference on Neural Networks.