Showing papers by "Romis Attux published in 2014"

Unorganized machines for seasonal streamflow series forecasting.

Abstract: Modern unorganized machines — extreme learning machines and echo state networks — provide an elegant balance between processing capability and mathematical simplicity, circumventing the difficultie...

A pragmatic entropy and differential entropy estimator for small datasets

Abstract: A pragmatic approach for entropy estimation is presented, first for discrete variables, then in the form of an extension for handling continuous and/or multivariate ones. It is based on coincidence detection, and its application leads to algorithms with three main attractive features: they are easy to use, can be employed without any a priori knowledge concerning source distribution (not even the alphabet cardinality K of discrete sources) and can provide useful estimates even when the number of samples is small (e.g. less than K, for discrete variables).

Multivariate PDF matching via kernel density estimation

Abstract: In this work, a measure of similarity based on the matching of multivariate probability density functions (PDFs) is proposed. In consonance with the information theoretic learning (ITL) framework, the affinity comparison between the joint PDFs is performed using a quadratic distance, estimated with the aid of the Parzen window method with Gaussian kernels. The motivation underlying this proposal is to introduce a criterion capable of quantifying, to a significant extent, the statistical dependence present on information sources endowed with temporal and/or spatial structure, like audio, images and coded data. The measure is analyzed and compared with the canonical ITL-based approach — correntropy — for a set of blind equalization scenarios. The comparison includes elements like surface analysis, performance comparison in terms of bit error rate and a qualitative discussion concerning image processing. It is also important to remark that the study includes the application of two computational intelligence paradigms: extreme learning machines and differential evolution. The results indicate that the proposal can be, in some scenarios, a more informative formulation than correntropy.

A Recurrence-Based Approach for Feature Extraction in Brain-Computer Interface Systems

Abstract: The feature extraction stage is one of the main tasks underlying pattern recognition, and, is particularly important for designing Brain-Computer Interfaces (BCIs), i.e. structures capable of mapping brain signals in commands for external devices. Within one of the most used BCIs paradigms, that based on Steady State Visual Evoked Potentials (SSVEP), such task is classically performed in the spectral domain, albeit it does not necessarily provide the best achievable performance. The aim of this work is to use recurrence-based measures in an attempt to improve the classification performance obtained with a classical spectral approaches for a five-command SSVEP-BCI system. For both recurrence and spectral spaces, features were selected using a cluster measure defined by the Davies-Bouldin index and the classification stage was based on linear discriminant analysis. As the main result, it was found that the threshold \(\varepsilon \) of the recurrence plot, chosen so as to yield a recurrence rate of 2.5 %, defined the key discriminant feature, typically providing a mean classification error of less than 2 % when information from 4 electrodes was used. Such classification performance was significantly better than that attained using spectral features, which strongly indicates that RQA is an efficient feature extraction technique for BCI.

cobICA: A concentration-based, immune-inspired algorithm for ICA over Galois fields

Abstract: An interesting and recent application of population-based metaheuristics resides in an unsupervised signal processing task: independent component analysis (ICA) over finite fields. Based on a state-of-the-art immune-inspired method, this work proposes a new ICA algorithm for finite fields of arbitrary order that employs mutation and local search operators specifically customized to the problem domain. The results obtained with the new technique indicate that the proposal is effective in performing component separation, and the analysis includes a preliminary study on image separation.

An implementation of SSVEP-BCI system based on a cluster measure for feature selection

Abstract: The main objective of a brain-computer interface (BCI) is to create alternative communication channels between the brain and a machine using information from cerebral responses. Among the possible paradigms to design a BCI system, this work focuses on Steady-State Visually Evoked Potentials (SSVEP). SSVEP are brain responses synchronized with fast repetitive external visual stimuli. The SSVEP-BCI system is able to meet many of the requirements of a strict BCI, but still needs to reduce the influence of noise on the Electroencephalogram (EEG) signal in order to improve its performance. In this paper, a novel SSVEP-BCI system is presented and analyzed in detail. The system is based on three pillars: spectrum estimation, systematic feature selection - for which different heuristics were proposed here -, and linear classification.

A comparative study of non-MSE criteria in nonlinear equalization

Abstract: This work studies the application of non-MSE criteria to adapt the linear readout of Extreme Learning Machines (ELMs) in the context of communication channel equalization. A qualitative and experimental analysis is performed, in terms of bit error rate, optimization surface and decision boundary. The results reached by the ELM-based equalizer, considering three different noise models, did not reveal clear advantages of using criteria based on the concepts of error entropy, correntropy, and the L 1 -norm of the error. Notwithstanding, the observed results motivate a theoretical investigation on the conditions under which the potential discrepancies between the optimal solutions of these criteria may be stressed.

A volterra filtering approach for the polynomial formulation of the constant modulus criterion

Abstract: In this work, an extended polynomial formulation of the constant modulus (CM) criterion under quadratic constraints is presented. Based on the method of Lagrange Multipliers, this ‘Volterra-CM formulation’ brings very relevant information about the structure of the null-gradient CM solutions in the equalizer parameter space, including a conjecture regarding the relationship between the smallest multiplier and the optimal CM receiver. In the case of a two-tap filter, the proposed formulation allows that the solutions be obtained in terms of a single parameter, the corresponding Lagrange multiplier. For filters with more than two taps, the problem requires that a nonlinear system be solved, which is done with the aid of an iterative algorithm. The obtained global convergence rates show that the formulation is an effective tool to describe the structure of the optimal CM solution.

Blind source separation for overdetermined linear quadratic mixtures of bandlimited signals

Abstract: In this paper, we address the problem of blind source separation for linear quadratic mixtures The proposed approach relies on the assumption that the input signals are band-limited As the nonlinearity of the mixing process tends to widen the spectra of the mixture signals, and taking into account the fact that there are more mixtures than sources in the overdetermined version of the problem, we propose a method that uses the additional mixtures to eliminate the nonlinearities of the observed signals This gives rise to a linear problem that can be solved with standard methods previously analyzed in the literature Numerical experiments show that the proposed algorithm successfully separates the sources under the proposed conditions

Analysis of the Weighted Fuzzy C-means in the Problem of Source Location

Abstract: This paper proposes the use of the clustering method called Weighted Fuzzy C-means to solve the problem of mixing matrix estimation in underdetermined source separation based on sparse component analysis. The performed comparative analysis shows that the approach has a significant application potential, especially if the distributions of the columns of the mixing matrix has a non-uniform character.

An MSE-based theoretical limit to the performance of linear source extraction and equalization methods in undermodeled scenarios

Abstract: This paper presents a simple and, to a certain extent, surprising result for Source Separation in an underdetermined scenario: without loss of generality, under the restriction that all sources have unit power, the sum of the residual mean-squared errors (MMSE) obtained after the estimation of all the sources is given by the difference between the number of sources and the number of sensors. This result can be extended to the case of single-input single-output (SISO) equalization, in which the obtained limit depends on the relationship between the length of the channel and equalizer impulse responses.