Kernel-based adaptive filters are sequential learning algorithms, operating on reproducing kernel Hilbert spaces. Their learning performance is susceptible to the selection of appropriate values for kernel bandwidth and learning-rate parameters. Additionally, as these algorithms train the model using a sequence of input vectors, their computation scales with the number of samples. We propose a framework that addresses the previous open challenges of kernel-based adaptive filters. In contrast to similar methods, our proposal sequentially optimizes the bandwidth and learning-rate parameters using stochastic gradient algorithms that maximize the correntropy function. To remove redundant samples, a sparsification approach based on dimensionality reduction is introduced. The framework is validated on both synthetic and real-world data sets. Results show that our proposal converges to relatively low values of mean-square-error while provides stable solutions in real-world applications.

The early detection of Alzheimer's disease and quantification of its progression poses multiple difficulties for machine learning algorithms. Two of the most relevant issues are related to missing data and results interpretability. To deal with both issues, we introduce a methodology to predict conversion of mild cognitive impairment patients to Alzheimer's from structural brain MRI volumes. First, we use morphological measures of each brain structure to build an instance-based feature mapping that copes with missed follow-up visits. Then, the extracted multiple feature mappings are combined into a single representation through the convex combination of reproducing kernels. The weighting parameters per structure are tuned based on the maximization of the centered-kernel alignment criterion. We evaluate the proposed methodology on a couple of well-known classification machines employing the ADNI database devoted to assessing the combined prognostic value of several AD biomarkers. The obtained experimental results show that our proposed method of Instance-based representation using multiple kernel learning enables detecting mild cognitive impairment as well as predicting conversion to Alzheimers disease within three years from the initial screening. Besides, the brain structures with larger combination weights are directly related to memory and cognitive functions.

In the recent past, estimating brain activity with magneto/electroencephalography (M/EEG) has been increasingly employed as a noninvasive technique for understanding the brain functions and neural dynamics. However, one of the main open problems when dealing with M/EEG data is its non-Gaussian and nonstationary structure. In this paper, we introduce a methodology for enhancing the data covariance estimation using a weighted combination of multiple Gaussian kernels, termed WM-MK, that relies on the Kullback-Leibler divergence for associating each kernel weight to its relevance. From the obtained results of validation on nonstationary and non-Gaussian brain activity (simulated and real-world EEG data), WM-MK proves that the accuracy of the source estimation raises by more effectively exploiting the measured nonlinear structures with high time and space complexity.

Accurate prediction of time series is of great interest because it can guide decisions in many economical or industrial fields. Methods can forecast either one or several steps ahead. The former is simpler and more common in many applications, while the latter is more challenging. The literature describes mainly three strategies of multi-step-ahead prediction: iterated (repeated one-step-ahead), direct, and MIMO (multiple input, multiple output). This paper proposes a MIMO strategy based on kernel adaptive filtering, which we named MSAKAF. The proposed MSAKAF has shown to be a more effective method in short, medium and long-term forecast. The proposed approach is validated on two real-world datasets. The results show that our proposal outperforms the compared baseline methods in terms of prediction accuracy.

We present a novel iterative regularized algorithm (IRA) for neural activity reconstruction that explicitly includes spatiotemporal constraints, performing a trade-off between space and time resolutions. For improving the spatial accuracy provided by electroencephalography (EEG) signals, we explore a basis set that describes the smooth, localized areas of potentially active brain regions. In turn, we enhance the time resolution by adding the Markovian assumption for brain activity estimation at each time period. Moreover, to deal with applications that have either distributed or localized neural activity, the spatiotemporal constraints are expressed through [Formula: see text] and [Formula: see text] norms, respectively. For the purpose of validation, we estimate the neural reconstruction performance in time and space separately. Experimental testing is carried out on artificial data, simulating stationary and non-stationary EEG signals. Also, validation is accomplished on two real-world databases, one holding Evoked Potentials and another with EEG data of focal epilepsy. Moreover, responses of functional magnetic resonance imaging for the former EEG data have been measured in advance, allowing to contrast our findings. Obtained results show that the [Formula: see text]-based IRA produces a spatial resolution that is comparable to the one achieved by some widely used sparse-based estimators of brain activity. At the same time, the [Formula: see text]-based IRA outperforms other similar smooth solutions, providing a spatial resolution that is lower than the sparse [Formula: see text]-based solution. As a result, the proposed IRA is a promising method for improving the accuracy of brain activity reconstruction.

In this paper, magnetic resonance image similarity metrics based on generative model induced spaces are introduced. Particularly, three generative-based similarities are proposed. Metrics are tested in an atlas selection task for multi-atlas-based image segmentation of basal ganglia structure, and compared with the mean square metric, as it is assessed on the high dimensional image domain. Attained results show that our proposal provides a suitable atlas selection and improves the segmentation of the structures of interest.

Systems that automatically detect voice pathologies are usually trained with recordings belonging to population of all ages. However such an approach might be inadequate because of the acoustic variations in the voice caused by the natural aging process. In top of that, elder voices present some perturbations in quality similar to those related to voice disorders, which make the detection of pathologies more troublesome. With this in mind, the study of methodologies which automatically incorporate information about speakers’ age, aiming at a simplification in the detection of voice disorders is of interest. In this respect, the present paper introduces an age detector trained with normal and pathological voice, constituting a first step towards the study of age-dependent pathology detectors. The proposed system employs sustained vowels of the Saarbrucken database from which two age groups are examinated: adults and elders. Mel frequency cepstral coefficients for characterization, and Gaussian mixture models for classification are utilized. In addition, fusion of vowels at score level is considered to improve detection performance. Results suggest that age might be effectively recognized using normal and pathological voices when using sustained vowels as acoustical material, opening up possibilities for the design of automatic age-dependent voice pathology detection systems.

A multichannel kernel adaptive filtering framework is presented that highlights relevant channels for the task of analyzing Motion Capture (MoCap) data. Functional relevance analysis is performed over input multichannel data by computing the pair-wise channel similarities to describe the main behavior of the considered applications. Particularly, the well-known Kernel Least Mean Square filter is enhanced using a correntropy-based similarity criterion between channel pairs. Besides, two sparseness criteria are studied to extract a sample subset that constructs a learning model displaying a good trade-off between filter complexity and accuracy. The proposed approach allows devising complex relationship among multi-channel time-series, revealing dependencies among the channels and the process time-structure. The method is tested in a well-known MoCap data set. Results show that our framework is an adequate alternative for finding functional relevance amongst multi-channel time-series.

Video surveillance systems are commonly designed based on background estimation and subtraction techniques to support further object detection and tracking procedures. In this paper, a new kernel-based adaptive learning approach, called Correntropy-based Adaptive Learning - CAL, is proposed to deal with stationary and non-stationary pixel dynamics to support video-based surveillance systems. To this end, an stochastic gradient based adaptive learning algorithm is enhanced using Correntropy cost function to learn, in a dynamic way, a background model that allows diminishing influence of intrinsic video artifacts. Moreover, automatic estimation of Correntropy kernel bandwidth value is carried out by means of shape variations of the CAL error distribution. Obtained results show that the CAL algorithm outperforms, in well-known real-world datasets, state of the art methodologies, suitably modeling the scene background with acceptable computational burden.

A new kernel-based image representation is proposed on this paper aiming to support clustering tasks on 3D magnetic resonances images. The approach establishes an effective way to encode inter-slice similarities, so that the main shape information is kept on a lower dimensional space. Additionally, a spectral clustering technique is employed to estimate a compact embedding space where natural groups are easily detectable. Proposed approach outperforms the conventional voxel-wise sum of squared differences on clustering the gender category. Additionally, a pair of eigenvectors describing accurately the subject age is found.

We propose a filtration approach to discriminate between stationary and non-stationary signals which consist into recursively update an enhanced representation of input time-series in such a way that the decomposition is able to identify time-varying statistical parameters of the data. The approach is based on the hypothesis that such updating providing a time-varying subspace projection under stationary constraints, allows to obtain a better separation. Validation of quality separation is carried on simulated and real data. In both cases, obtained separation shows that proposed approach is able to identify different dynamics on analyzed data.

In this paper we propose a kernel spectral clustering-based technique to catch the different regimes experienced by a time-varying system. Our method is based on a multiple kernel learning approach, which is a linear combination of kernels. The calculation of the linear combination coefficients is done by determining a ranking vector that quantifies the overall dynamical behavior of the analyzed data sequence over-time. This vector can be calculated from the eigenvectors provided by the the solution of the kernel spectral clustering problem. We apply the proposed technique to a trial from the Graphics Lab Motion Capture Database from Carnegie Mellon University, as well as to a synthetic example, namely three moving Gaussian clouds. For comparison purposes, some conventional spectral clustering techniques are also considered, namely, kernel k-means and min-cuts. Also, standard k-means. The normalized mutual information and adjusted random index metrics are used to quantify the clustering performance. Results show the usefulness of proposed technique to track dynamic data, even being able to detect hidden objects.

BACKGROUND: Proteins are the key elements on the path from genetic information to the development of life. The roles played by the different proteins are difficult to uncover experimentally as this process involves complex procedures such as genetic modifications, injection of fluorescent proteins, gene knock-out methods and others. The knowledge learned from each protein is usually annotated in databases through different methods such as the proposed by The Gene Ontology (GO) consortium. Different methods have been proposed in order to predict GO terms from primary structure information, but very few are available for large-scale functional annotation of plants, and reported success rates are much less than the reported by other non-plant predictors. This paper explores the predictability of GO annotations on proteins belonging to the Embryophyta group from a set of features extracted solely from their primary amino acid sequence. RESULTS: High predictability of several GO terms was found for Molecular Function and Cellular Component. As expected, a lower degree of predictability was found on Biological Process ontology annotations, although a few biological processes were easily predicted. Proteins related to transport and transcription were particularly well predicted from primary structure information. The most discriminant features for prediction were those related to electric charges of the amino-acid sequence and hydropathicity derived features. CONCLUSIONS: An analysis of GO-slim terms predictability in plants was carried out, in order to determine single categories or groups of functions that are most related with primary structure information. For each highly predictable GO term, the responsible features of such successfulness were identified and discussed. In addition to most published studies, focused on few categories or single ontologies, results in this paper comprise a complete landscape of GO predictability from primary structure encompassing 75 GO terms at molecular, cellular and phenotypical level. Thus, it provides a valuable guide for researchers interested on further advances in protein function prediction on Embryophyta plants.

Automatic detection of hypernasality in voices of children with Cleft Lip and Palate (CLP) is made considering two charcaterization techniques, one based on acoustic, noise and cepstral analysis and other based on nonlinear dynamic features. Besides characterization, two automatic feature selection techniques are implemented in order to find optimal sub-spaces to better discriminate between healthy and hypernasal voices. Results indicate that nonlinear dynamic features are valuable tool for automatic detection of hypernasality; addtionally both feature selection techniques show stable and consistent results, achieving accuracy levels of up to 93.73%. Index Terms: Hypernasality, Cleft Lip and Palate, acoustic, cepstral, nonlinear dynamics.

An effective data representation methodology on high-dimension feature spaces is presented, which allows a better interpretation of subjacent physiological phenomena (namely, cardiac behavior related to cardiovascular diseases), and is based on search criteria over a feature set resulting in an increase in the detection capability of ischemic pathologies, but also connecting these features with the physiologic representation of the ECG. The proposed dimension reduction scheme consists of three levels: projection, interpretation, and visualization. First, a hybrid algorithm is described that projects the multidimensional data to a lower dimension space, gathering the features that contribute similarly in the meaning of the covariance reconstruction in order to find information of clinical relevance over the initial training space. Next, an algorithm of variable selection is provided that further reduces the dimension, taking into account only the variables that offer greater class separability, and finally, the selected feature set is projected to a 2-D space in order to verify the performance of the suggested dimension reduction algorithm in terms of the discrimination capability for ischemia detection. The ECG recordings used in this study are from the European ST-T database and from the Universidad Nacional de Colombia database. In both cases, over 99% feature reduction was obtained, and classification precision was over 99% using a five-nearest-neighbor classifier (5-NN).