Browsing by Author "Teixeira, A.R."
Now showing 1 - 3 of 3
Results Per Page
Sort Options
- Single-channel electroencephalogram analysis using non-linear subspace techniquesPublication . Teixeira, A.R.; Alves, N.; Tome, A.M.; Bohm, M.; Lang, E.W.; Puntonet, C.G.In this work, we propose the correction of univariate, single channel EEGs using projective subspace techniques. The biomedical signals which often represent one dimensional time series, need to be transformed to multi-dimensional signal vectors for the latter techniques to be applicable. The transformation can be achieved by embedding an observed signal in its delayed coordinates. We propose the application of two non-linear subspace techniques to the obtained multidimensional signal. One of the techniques consists in a modified version of Singular Spectrum Analysis (SSA) and the other is kernel Principal Component Analysis (KPCA) implemented using a reduced rank approximation of the kernel matrix. Both nonlinear subspace projection techniques are applied to an electroencephalogram (EEG) signal recorded in the frontal channel to extract its prominent electrooculogram (EOG) interference.
- SSA of biomedical signals: A linear invariant systems approachPublication . Figueiredo, N.; Georgieva, P.; Lang, E.W.; Santos, I.M.; Teixeira, A.R.; Tomé, A.M.Singular spectrum analysis (SSA) is considered from a linear invariant systems perspective. In this terminology, the extracted components are considered as outputs of a linear invariant system which corresponds to finite impulse response (FIR) filters. The number of filters is determined by the embedding dimension.We propose to explicitly define the frequency response of each filter responsible for the selection of informative components. We also introduce a subspace distance measure for clustering subspace models. We illustrate the methodology by analyzing Electroencephalograms (EEG).
- Unsupervised feature extraction via kernel subspace techniquesPublication . Teixeira, A.R.; Tomé, A.M.; Lang, E.W.This paper provides a new insight into unsupervised feature extraction techniques based on kernel subspace models. The data projected onto kernel subspace models are new data representations which might be better suited for classification. The kernel subspace models are always described exploiting the dual form for the basis vectors which requires that the training data must be available even during the test phase. By exploiting an incomplete Cholesky decomposition of the kernel matrix, a computationally less demanding implementation is proposed. Online benchmark data sets allow the evaluation of these feature extraction methods comparing the performance of two classifiers which both have as input either the raw data or the new representations.