Papers by Manel Martínez-ramón
NeuroImage, 2011
Pattern classification of brain imaging data can enable the automatic detection of differences in... more Pattern classification of brain imaging data can enable the automatic detection of differences in cognitive processes of specific groups of interest. Furthermore, it can also give neuroanatomical information related to the regions of the brain that are most relevant to detect these differences by means of feature selection procedures, which are also well-suited to deal with the high dimensionality of brain imaging data. This work proposes the application of recursive feature elimination using a machine learning algorithm based on composite kernels to the classification of healthy controls and patients with schizophrenia. This framework, which evaluates nonlinear relationships between voxels, analyzes whole-brain fMRI data from an auditory task experiment that is segmented into anatomical regions and recursively eliminates the uninformative ones based on their relevance estimates, thus yielding the set of most discriminative brain areas for group classification. The collected data was processed using two analysis methods: the general linear model (GLM) and independent component analysis (ICA). GLM spatial maps as well as ICA temporal lobe and default mode component maps were then input to the classifier. A mean classification accuracy of up to 95% estimated with a leave-two-out cross-validation procedure was achieved by doing multi-source data classification. In addition, it is shown that the classification accuracy rate obtained by using multi-source data surpasses that reached by using single-source data, hence showing that this algorithm takes advantage of the complimentary nature of GLM and ICA.
Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 2014
This paper provides an overview of the support vector machine (SVM) methodology and its applicabi... more This paper provides an overview of the support vector machine (SVM) methodology and its applicability to real-world engineering problems. Specifically, the aim is to review the current state of the SVM technique, and to show some of its latest successful results in real problems 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 present different case studies in real engineering problems, where the application of the SVM methodology has obtained 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 real results on this topic 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: 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.
Support Vector Machines for Antenna Array Processing and Electromagnetics
Synthesis Lectures on Computational Electromagnetics, 2006
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system,... more All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any meanselectronic, mechanical, photocopy, recording, or any other except for brief quotations in printed reviews, without the prior ...
Detecting failure of antenna array elements using machine learning optimization
2007 IEEE Antennas and Propagation International Symposium, 2007
A Multi-class support vector classifier (SVC) is proposed for planar array failure diagnosis. Ext... more A Multi-class support vector classifier (SVC) is proposed for planar array failure diagnosis. Extracted feature information from the far field intensity of the array is used to train and test the multi-class SVC, so one can detect the location of failed elements in an array and also the level of failure.
2009 SBMO/IEEE MTT-S International Microwave and Optoelectronics Conference (IMOC), 2009
The estimation of the spectrum usage from the point of view of number of users and modulation typ... more The estimation of the spectrum usage from the point of view of number of users and modulation types is addressed in this paper. The techniques used here are based on Support Vector Machines (SVM). SVMs are machine learning strategies which use a robust cost function alternative to the widely used Least Squares function and that apply a regularization which provides control of the complexity of the resulting estimators. As a result, estimators are robust against interferences and nongaussian noise and present excellent generalization properties where the number of data available for the estimation is small. The structure pre sented here has a feature extraction part that, instead of using an FFT approach, uses the SVM criterion for spectrum estimation, feature extraction and modulation classification .
For least mean-square (LMS) algorithm applications, it is important to improve the speed of conve... more 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 leastsquares (RLS) filters. r
École Polytechnique Fédérale de Lausanne, Lausanne 1015, Switzerland
Robust-filter using support vector machines
Despite its reliable diagnosis, schizophrenia lacks an objective diagnostic test or a validated b... more Despite its reliable diagnosis, schizophrenia lacks an objective diagnostic test or a validated biomarker, which prevents a better understanding of this disorder. Structural magnetic resonance imaging (sMRI) has been vastly explored to find consistent abnormality patterns of gray matter concentration (GMC) in schizophrenia, yet we are far from having reached conclusive evidence. This paper presents a machine learning approach based on resampling techniques to find brain regions with consistent patterns of GMC differences between healthy controls and schizophrenia patients, these regions being detected by means of source-based morphometry. This work uses multi-site data from the Mind Clinical Imaging Consortium, which is composed of sMRI data from 124 controls and 110 patients. Our method achieves a better classification rate than other algorithms and detects regions with GMC differences between both groups that are consistent with several findings on the literature. In addition, the results obtained on data from multiple sites suggest that it may be possible to replicate these results on other datasets.
An Introduction to Kernel Methods
This paper presents a review in the form of a unified framework for tackling estimation problems ... more This paper presents a review in the form of a unified framework for tackling estimation problems in Digital Signal Processing (DSP) using Support Vector Machines (SVMs). The paper formalizes our developments in the area of DSP with SVM principles. The use of SVMs for DSP is already mature, and has gained popularity in recent years due to its advantages over other methods: SVMs are flexible non-linear methods that are intrinsically regularized and work well in low-sample-sized and high-dimensional problems. SVMs can be designed to take into account different noise sources in the formulation and to fuse heterogeneous information sources. Nevertheless, the use of SVMs in estimation problems has been traditionally limited to its mere use as a black-box model. Noting such limitations in the literature, we take advantage of several properties of Mercer's kernels and functional analysis to develop a family of SVM methods for estimation in DSP. Three types of signal model equations are analyzed. First, when a specific time-signal structure is assumed to model the underlying system that generated the data, the linear signal model (so called Primal Signal Model formulation) is first stated and analyzed. Then, non-linear versions of the signal structure can be readily developed by following two different approaches. On the one hand, the signal model equation is written in reproducing kernel Hilbert spaces (RKHS) using the well-known RKHS Signal Model formulation, and Mercer's kernels are readily used in SVM nonlinear algorithms. On the other hand, in the alternative and not so common Dual Signal Model formulation, a signal expansion is made by using an auxiliary signal model equation given by a non-linear regression of each time instant in the observed time series. These building blocks can be used to generate different novel SVM-based methods for problems of signal estimation, and we deal with several of the most important ones in DSP. We illustrate the usefulness of this methodology by defining SVM algorithms for linear and non-linear system identification, spectral analysis, nonuniform interpolation, sparse deconvolution, and array processing. The performance of the developed SVM methods is compared to standard approaches in all these settings. The experimental results illustrate the generality, simplicity, and capabilities of the proposed SVM framework for DSP.
Spectrally adapted Mercer kernels for support vector nonuniform interpolation
ABSTRACT Interpolation of nonuniformly sampled signals in the presence of noise is a widely analy... more ABSTRACT Interpolation of nonuniformly sampled signals in the presence of noise is a widely analyzed problem in signal processing applications. Interpolators based on Support Vector Machines (SVM) with Gaussian and sinc Mercer kernels have been previously proposed, obtaining good performance in terms of regularization and sparseness. In this paper, inspired in the classical spectral interpretation of the Wiener filter, we explore the impact of adapting the spectrum of the SVM kernel to that of the observed signal. We provide a theoretical foundation for this approach based on a continuous-time equivalent system for interpolation. We study several kernels with different degrees of spectral adaptation to band-pass signals, namely, modulated kernels and autocorrelation kernels. The proposed algorithms are evaluated with extensive simulations with synthetic signals and an application example with real data. Our approach is compared with SVM with Gaussian and sinc kernels and with other well known interpolators. The SVM with autocorrelation kernel provides the highest performance in terms of signal to error ratio in several scenarios. We conclude that the estimated (or actual if known) autocorrelation of the observed sequence can be straightforwardly used as a spectrally adapted kernel, outperforming the classic SVM with low pass kernels for nonuniform interpolation.
Lecture Notes in Computer Science, 2002
A new approach to the non-parametric spectral estimation on the basis of the Support Vector (SV) ... more A new approach to the non-parametric spectral estimation on the basis of the Support Vector (SV) framework is presented. Two algorithms are derived for both uniform and non-uniform sampling. The relationship between the SV free parameters and the underlying process statistics is discussed. The application in two real data examples, the sunspot numbers and the Heart Rate Variability, shows the higher resolution and robustness in SV spectral analysis algorithms.
Neurocomputing, 2009
A family of kernel methods, based on the g-filter structure, is presented for non-linear system
Voxel Selection in MRI through Bagging and Conformal Analysis: Application to Detection of Obsessive Compulsive Disorder
2012 Second International Workshop on Pattern Recognition in NeuroImaging, 2012
ABSTRACT In this work we apply a multivariate feature selection method based on bagging linear SV... more ABSTRACT In this work we apply a multivariate feature selection method based on bagging linear SVMs to construct a classifier able to differentiate among control subjects and patients with obsessive compulsive disorder (OCD). Our method selects sets of voxels that are relevant for the detection of the disease. The voxel selection is completed with a conformal analysis based refinement that controls over fitting and dramatically reduces the test error rate of the final classifier. Furthermore, the resulting discrimination pattern is organized in regions that show great agreement with those found by traditional methods used in OCD problems, achieving cleaner and more accurate region maps.
Identification of OCD-Relevant Brain Areas through Multivariate Feature Selection
Lecture Notes in Computer Science, 2012
IEEE Signal Processing Letters, 2006
A new support vector machine (SVM) algorithm for coherent robust demodulation in orthogonal frequ... more A new support vector machine (SVM) algorithm for coherent robust demodulation in orthogonal frequency-division multiplexing (OFDM) systems is proposed. We present a complex regression SVM formulation specifically adapted to a pilots-based OFDM signal. This novel proposal provides a simpler scheme than an SVM classification method. The feasibility of our approach is substantiated by computer simulation results obtained for IEEE 802.16
Abstract—We introduce two support vector machine (SVM)- based approaches for solving antenna prob... more Abstract—We introduce two support vector machine (SVM)- based approaches for solving antenna problems such as beamforming, sidelobe suppression, and maximization of the signal-to-noise ratio. A basic introduction to SVM optimization is provided and a complex nonlinear SVM formulation developed to handle antenna array processing in space and time. The new opti- mization formulation is compared with both the minimum mean
Explicit Recursive and Adaptive Filtering in Reproducing Kernel Hilbert Spaces
IEEE Transactions on Neural Networks and Learning Systems, 2014
ABSTRACT This brief presents a methodology to develop recursive filters in reproducing kernel Hil... more ABSTRACT This brief presents a methodology to develop recursive filters in reproducing kernel Hilbert spaces. Unlike previous approaches that exploit the kernel trick on filtered and then mapped samples, we explicitly define the model recursivity in the Hilbert space. For that, we exploit some properties of functional analysis and recursive computation of dot products without the need of preimaging or a training dataset. We illustrate the feasibility of the methodology in the particular case of the gamma-filter, which is an infinite impulse response filter with controlled stability and memory depth. Different algorithmic formulations emerge from the signal model. Experiments in chaotic and electroencephalographic time series prediction, complex nonlinear system identification, and adaptive antenna array processing demonstrate the potential of the approach for scenarios where recursivity and nonlinearity have to be readily combined.
Uploads
Papers by Manel Martínez-ramón