The multitemporal classification of remote sensing images is a challenging problem, in which the efficient combination of different sources of information (e.g., temporal, contextual, or multisensor) can improve the results. In this paper, we present a general framework based on kernel methods for the integration of heterogeneous sources of information. Using the theoretical principles in this framework, three main contributions are presented. First, a novel family of kernel-based methods for multitemporal classification of remote sensing images is presented. The second contribution is the development of nonlinear kernel classifiers for the well-known difference and ratioing change detection methods by formulating them in an adequate high-dimensional feature space. Finally, the presented methodology allows the integration of contextual information and multisensor images with different levels of nonlinear sophistication. The binary support vector (SV) classifier and the one-class SV domain description classifier are evaluated by using both linear and nonlinear kernel functions. Good performance on synthetic and real multitemporal classification scenarios illustrates the generalization of the framework and the capabilities of the proposed algorithms.

This paper deals with the application of the Support Vector Method (SVM) methodology to the Auto Regressive and Moving Average (ARMA) linear-system identification problem. The SVM-ARMA algorithm for a single-input single-output transfer function is formulated. The relationship between the SVM coefficients and the residuals, together with the embedded estimation of the autocorrelation function, are presented. Also, the effect of the numerical regularization is used to highlight the robust cost character of this approach. A clinical example is presented for qualitative comparison with the classical Least Squares (LS) methods.

This paper provides an overview of the support vector machine (SVM) methodology and its applicability to real‐world engineering problems. Specifically, the aim of this study is to review the current state of the SVM technique, and to show some of its latest successful results in real‐world problems present in different engineering fields. The paper starts by reviewing the main basic concepts of SVMs and kernel methods. Kernel theory, SVMs, support vector regression (SVR), and SVM in signal processing and hybridization of SVMs with meta‐heuristics are fully described in the first part of this paper. The adoption of SVMs in engineering is nowadays a fact. As we illustrate in this paper, SVMs can handle high‐dimensional, heterogeneous and scarcely labeled datasets very efficiently, and it can be also successfully tailored to particular applications. The second part of this review is devoted to different case studies in engineering problems, where the application of the SVM methodology has led to excellent results. First, we discuss the application of SVR algorithms in two renewable energy problems: the wind speed prediction from measurements in neighbor stations and the wind speed reconstruction using synoptic‐pressure data. The application of SVMs in noninvasive cardiac indices estimation is described next, and results obtained there are presented. The application of SVMs in problems of functional magnetic resonance imaging (fMRI) data processing is further discussed in the paper: brain decoding and mental disorder characterization. The following application deals with antenna array processing, namely SVMs for spatial nonlinear beamforming, and the SVM application in a problem of arrival angle detection. Finally, the application of SVMs to remote sensing image classification and target detection problems closes this review. WIREs Data Mining Knowl Discov 2014, 4:234–267. doi: 10.1002/widm.1125 This article is categorized under: Technologies > Computational Intelligence Technologies > Machine Learning

Abstract-Nonlinear system identification based on support vector machines (SVM) has been usually addressed by means of the standard SVM regression (SVR), which can be seen as an implicit nonlinear autoregressive and moving average (ARMA) model in some reproducing kernel Hilbert space (RKHS). The proposal of this letter is twofold. First, the explicit consideration of an ARMA model in an RKHS (SVM-ARMA ) is proposed. We show that stating the ARMA equations in an RKHS leads to solving the regularized normal equations in that RKHS, in terms of the autocorrelation and cross correlation of the (nonlinearly) transformed input and output discrete time processes. Second, a general class of SVM-based system identification nonlinear models is presented, based on the use of composite Mercer's kernels. This general class can improve model flexibility by emphasizing the input-output cross information (SVM-ARMA ), which leads to straightforward and natural combinations of implicit and explicit ARMA models (SVR-ARMA and SVR-ARMA ). Capabilities of these different SVM-based system identification schemes are illustrated with two benchmark problems.

Pattern classification in functional MRI (fMRI) is a novel methodology to automatically identify differences in distributed neural substrates resulting from cognitive tasks. Reliable pattern classification is challenging due to the high dimensionality of fMRI data, the small number of available data sets, interindividual differences, and dependence on the acquisition methodology. Thus, most previous fMRI classification methods were applied in individual subjects. In this study, we developed a novel approach to improve multiclass classification across groups of subjects, field strengths, and fMRI methods. Spatially normalized activation maps were segmented into functional areas using a neuroanatomical atlas and each map was classified separately using local classifiers. A single multiclass output was applied using a weighted aggregation of the classifier's outputs. An Adaboost technique was applied, modified to find the optimal aggregation of a set of spatially distributed classifiers. This Adaboost combined the regionspecific classifiers to achieve improved classification accuracy with respect to conventional techniques. Multiclass classification accuracy was assessed in an fMRI group study with interleaved motor, visual, auditory, and cognitive task design. Data were acquired across 18 subjects at different field strengths (1.5 T, 4 T), with different pulse sequence parameters (voxel size and readout bandwidth). Misclassification rates of the boosted classifier were between 3.5% and 10%, whereas for the single classifier, these were between 15% and 23%, suggesting that the boosted classifier provides a better generalization ability together with better robustness. The high computational speed of boosting classification makes it attractive for real-time fMRI to facilitate online interpretation of dynamically changing activation patterns.Keywords: Functional magnetic resonance imaging; Pattern classification; Support vector machines; Adaboost IntroductionBrain activation changes in response to even simple sensory input and motor tasks encompass a widely distributed network of functional brain areas. Information embedded in the spatial shape and extent of these activation patterns, and differences in voxel-tovoxel time course, are not easily quantified with conventional analysis tools, such as statistical parametric mapping (SPM) (Kiebel and Friston, 2004a,b). Pattern classification in functional MRI (fMRI) is a novel approach, which promises to characterize subtle differences in activation patterns between different tasks. However, automatic and reliable classification of patterns is challenging due to the high dimensionality of fMRI data, the small number of available data sets, interindividual differences in activation patterns, and dependence on the image acquisition methodology. Recent work by Cox and Savoy (2003) demonstrated that linear discriminant analysis and support vector machines (SVM) allow 10-way discrimination of visual activation patterns evoked by the visual presentation of various categories of ...

For least mean square (LMS) algorithm applications, it is important to improve the speed of convergence vs the residual error trade off imposed by the selection of a certain value for the step size. In this paper, we propose to use a mixture approach, adaptively combining two independent LMS filters with large and small step sizes to obtain fast convergence with low misadjustment during stationary periods. Some plant identification simulation examples show the effectiveness of our method when compared to previous variable step size approaches. This combination approach can be straightforwardly extended to other kinds of filters, as it is illustrated with a convex combination of recursive least squares (RLS) filters..

No abstract

No abstract