SEARCH · PubMed Health

Results for “Gaussian process”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 739 records · Page 41Linked to original sources

PubMed2003

Prediction of drug release profiles using an intelligent learning system: an experimental study in transdermal iontophoresis.

This paper investigates the use of a neural-network-based intelligent learning system for the prediction of drug release profiles. An experimental study in transdermal iontophoresis (TI) is employed to evaluate the applicability of a particular neural network (NN) model, i.e. the Gaussian mixture model (GMM), in modeling and predicting drug release profiles. A number of tests are systematically designed using the face-centered central composite design (CCD) approach to examine the effects of various process variables simultaneously during the iontophoresis process. The GMM is then applied to model and predict the drug release profiles based on the data samples collected from the experiments. The GMM results are compared with those from multiple regression models. In addition, the bootstrap method is used to assess the reliability of the network predictions by estimating confidence intervals associated with the results. The results demonstrate that the combination of the face-centered CCD and GMM can be employed as a useful intelligent tool for the prediction of time-series profiles in pharmaceutical and biomedical experiments.

Administration, Cutaneous↗

PubMed2002

New Bayesian discriminator for detection of atrial tachyarrhythmias.

BACKGROUND: Accurate, rapid detection of atrial tachyarrhythmias has important implications in the use of implantable devices for treatment of cardiac arrhythmia. Currently available detection algorithms for atrial tachyarrhythmias, which use the single-index method, have limited sensitivity and specificity. METHODS AND RESULTS: In this study, we evaluated the performance of a new Bayesian discriminator algorithm in the detection of atrial fibrillation (AF), atrial flutter (AFL), and sinus rhythm (SR). Bipolar recording of 364 rhythms (AF=156, AFL=88, SR=120) at the high right atrium were collected from 20 patients who underwent electrophysiological procedures. After initial signal processing, a column vector of 5 features for each rhythm were established, based on the regularity, rate, energy distribution, percent time of quiet interval, and baseline reaching of the rectified autocorrelation coefficient functions. Rhythm identification was obtained by use of Bayes decision rule and assumption of Gaussian distribution. For the new Bayesian discriminator, the overall sensitivity for detection of SR, AF, and AFL was 97%, 97%, and 94%, respectively; and the overall specificity for detection of SR, AF, and AFL was 98%, 98%, and 99%, respectively. The overall accuracy of detection of SR, AF, and AFL was 98%, 97% and 98%, respectively. Furthermore, sensitivity, specificity, and accuracy of this algorithm were not affected by a range of white Gaussian noises with different intensities. CONCLUSIONS: This new Bayesian discriminator algorithm, based on Bayes decision of multiple features of atrial electrograms, allows rapid on-line and accurate (98%) detection of AF with robust anti-noise performance.

Adolescent↗

PubMed2003

Application of wavelet transforms and an approximate deconvolution method for the resolution of noisy overlapped peaks in DNA capillary electrophoresis.

A new procedure for resolving noisy overlapped peaks in DNA separations by capillary electrophoresis (CE) is developed. The procedure combines both a wavelet-based denoising method that effectively denoises the signal and a novel approximate deconvolution technique that resolves the fragment peaks and improves the ability to separate highly overlapped peaks early in the electrophoresis process. Different kinds of overlapped peaks with and without noise simulated by computer as well as some DNA experimental electropherograms were submitted to the new procedure. A second order differential operator with variable coefficients is applied to the entire electrophoresis signal at any given time and approximate deconvolutions of the individual Gaussian peaks are performed. The operator incorporates the effect of the superposition and gives exact annihilation in the neighborhood of each peak. Overlapped peaks with a resolution higher than 0.46 can be resolved directly. Also, the method can determine the peak components of signals with a signal to noise ratio higher than 1.4

DNA Fragmentation↗

PubMed1996

Application of robust data processing methods to the analysis of eye movements.

We have developed a robust, nonlinear differentiating digital filter for the estimation of eye velocity from an eye-position signal. This filter is equally applicable to a large variety of other biomedical signals. The filter is implemented by taking an odd number of two-point differences around the point of interest and then selecting the median difference. Dividing the median difference by the time interval yields the derivative of the input signal. The Robust Differentiator (RD) is a classic order-statistic filter consisting of a bank of linear filters (the two-point differences) followed by the median operation to select one difference for the derivative calculation. The RD has no impulse response function and eliminates the "ringing" that is typical of all linear filters. The RD not only performs better than the FIR differentiating filters for impulse noise but is equally effective for both broad-band and narrow-band Gaussian noise. The RD "bandwidth" is adjustable by selecting the interval for the two-point differences.

Eye Movements↗

PubMed2004

Dynamic monitoring system for full-scale wastewater treatment plants.

This paper proposes a new process monitoring method using dynamic independent component analysis (ICA), ICA is a recently developed technique to extract the hidden factors that underlie sets of measurements, whereas principal component analysis (PCA) is a dimensionality reduction technique in terms of capturing the variance of the data. Its goal is to find a linear representation of non-Gaussian data so that the components are statistically independent. PCA aims at finding PCs that are uncorrelated and are linear combinations of the observed variables, while ICA is designed to separate the ICs that are independent and constitute the observed variables. The dynamic ICA monitoring method is applying ICA to the augmenting matrix with time-lagged variables. The dynamic monitoring method was applied to detect and monitor disturbances in a full-scale biological wastewater treatment (WWTP), which is characterized by a variety of dynamic and non-Gaussian characteristics. The dynamic ICA method showed more powerful monitoring performance on a WWTP application than the dynamic PCA method since it can extract source signals which are independent of time and cross-correlation of variables.

Algorithms↗

PubMed1992

Development of a digital adaptive control system for PO2 regulation in a membrane oxygenator.

Regulation of gas exchange in artificial lungs (oxygenators) during cardiopulmonary bypass is normally achieved by manual control of the gas composition and flow in response to intermittent sampling of the arterial partial pressures of oxygen (PaO2) and carbon dioxide (PaCO2). Manual control often results in abnormal blood gases which have been implicated in patient morbidity as well as influencing perfusion safety. Fine control of PaO2 and PaCO2 may be achieved by a combination of an in-line blood gas monitoring system and a membrane type oxygenator which is automatically regulated. The overall dynamics of the oxygenation process and control system components are complex and have nonlinear, multivariable and time-varying characteristics. Consequently, an adaptive control system approach is necessary. The implementation of a digital self-tuning control regime for PaO2 is described here. The controller is based on an explicit Linear Quadratic Gaussian (LQG) self-tuning control design which is presented using a polynomial equation approach. The controller performance was investigated in in vitro experiments. The self-tuner performed satisfactorily with various sensor/oxygenator combinations for blood flow and temperature load disturbances. In contrast, a nonadaptive (proportional-integral, PI) type of control system was found to be unsuitable.

Carbon Dioxide↗

PubMed2001

High-performance liquid chromatography analysis of spectrin oligomerization.

Gel filtration chromatography has been used to analyze the oligomerization of human erythrocyte spectrin. By applying an exponentially modified Gaussian function we have been able to resolve overlapping elution peaks. From these peaks it was possible to calculate the equilibrium composition of each spectrin concentration and thus also the dissociation constants describing the oligomeric process. The determined dissociation constants for tetramer formation (1.3 microM) and for hexamer formation (24 microM) agree well with other measurements.

Chromatography, Gel↗

PubMed2003

Analysis of quantization error in high-precision continuous-flow isotope ratio mass spectrometry.

High-precision isotope ratio mass spectrometry (IRMS) systems are equipped with digitizers that deliver effective maximum digitization depths of 16 to 24 bits; however, there are no analyses of the proper board depth required to retain high precision in continuous-flow techniques. We report an experimental and theoretical evaluation of quantization error in continuous-flow IRMS (CF-IRMS). CO2 samples (100 pmol-30 nmol) were injected into a gas chromatography combustion IRMS system (GC-CIRMS). The analog signal was digitized by high precision, 24-bit ADC boards at 10 Hz, and was post-processed to simulate 12, 14, and 16-bit data sets. Delta13Cpdh values were calculated for all data sets by the conventional "summation" method or by curve-fitting the chromatographic peaks to the exponentially modified Gaussian (EMG) function. Benchmarks of S.D.(delta13Cpdh) = 0.3, 0.6, and 1.0/1000 were considered to assess precision. In the presence of significant quantization noise, curve-fitting required several-fold less CO2 than the summation method to reach a given benchmark. We derived an equation to describe the theoretical limitations of precision for the summation method as a function of CO2 admitted to the source and the step size of the boards. Theory was in close agreement with the observed lower limit of precision for the simulated 16-bit data set. Curve-fitting achieved a precision of S.D. <0.3/1000 for injections 20-fold smaller than summation for CO2 samples collected on an IRMS with 16-bit resolution. By mitigating the impact of quantization noise, curve-fitting expands the dynamic range within a single run to include lower analyte levels, and effectively reduces the need for high pumping capacities and high precision ADC boards.

Isotopes↗

PubMed2004

Comparison between the predictions of a Gaussian plume model and a Lagrangian particle dispersion model for annual average calculations of long-range dispersion of radionuclides.

This study attempts to validate the applicability of a simple Gaussian dispersion model for predicting long-range dispersion of continuous releases from an industrial site, by comparison with a Lagrangian particle dispersion model. The United Kingdom Meteorological Office model NAME has been used to predict annual average concentrations of radionuclides over Western Europe, resulting from discharges from the British Nuclear Fuels (BNFL) site at Sellafield, UK. The results are compared here to calculations performed using a conventional Gaussian type of dispersion model, PLUME. The results of the Gaussian model were compared at 14 locations within Western Europe at long range (up to 1700 km). The differences in predictions between the models were explained readily by differences in the way dispersion and deposition processes are represented in the two models. However, differences are generally small compared to the expected precision of the models. The implementation of environmental processes in NAME is more complete and realistic than in PLUME, and as such the results from this model may be considered more realistic. However, given that PLUME is much simpler to use, and appears to over-estimate, rather than under-estimate, environmental concentrations, its use for radiological assessments appears appropriate.

Air Movements↗

PubMed2002

Gene expression data analysis of human lymphoma using support vector machines and output coding ensembles.

The large amount of data generated by DNA microarrays was originally analysed using unsupervised methods, such as clustering or self-organizing maps. Recently supervised methods such as decision trees, dot-product support vector machines (SVM) and multi-layer perceptrons (MLP) have been applied in order to classify normal and tumoural tissues. We propose methods based on non-linear SVM with polynomial and Gaussian kernels, and output coding (OC) ensembles of learning machines to separate normal from malignant tissues, to classify different types of lymphoma and to analyse the role of sets of coordinately expressed genes in carcinogenic processes of lymphoid tissues. Using gene expression data from "Lymphochip", a specialised DNA microarray developed at Stanford University School of Medicine, we show that SVM can correctly separate normal from tumoural tissues, and OC ensembles can be successfully used to classify different types of lymphoma. Moreover, we identify a group of coordinately expressed genes related to the separation of two distinct subgroups inside diffuse large B-cell lymphoma (DLBCL), validating a previous Alizadeh's hypothesis about the existence of two distinct diseases inside DLBCL.

Artificial Intelligence↗

PubMed2004

Generalized randomly amplified linear system driven by gaussian noises: extreme heavy tail and algebraic correlation decay in plasma turbulence.

The extreme heavy tail and the power-law decay of the turbulent flux correlation observed in hot magnetically confined plasmas are modeled by a system of coupled Langevin equations describing a continuous time linear randomly amplified stochastic process where the amplification factor is driven by a superposition of colored noises which, in a suitable limit, generate a fractional Brownian motion. An exact analytical formula for the power-law tail exponent beta is derived. The extremely small value of the heavy tail exponent and the power-law distribution of laminar times also found experimentally are obtained, in a robust manner, for a wide range of input values, as a consequence of the (asymptotic) self-similarity property of the noise spectrum. As a by-product, a new representation of the persistent fractional Brownian motion is obtained.

Journal Article↗

PubMed1996

A stereo model using LoG and Gabor filtering.

This paper presents a computational model for stereopsis. Laplacian of Gaussian filters are used to extract spatial features and a set of one-octave Gabor filters is used to extract orientation information from a scene. A mapping from a 3D space onto two 2D image planes is derived using a Gaussian sphere model. The determinant of the Jacobian of the mapping is derived and matching is performed using zero-crossings associated with their orientation information. Relaxation labelling is used as a co-operative process for solving the correspondence problem.

Computer Simulation↗

PubMed2003

Resolution improvement with dispersion manipulation and a retrieval algorithm in optical coherence tomography.

We propose and demonstrate what is to our knowledge a novel technique of improving the spatial resolution of an optical coherence tomography (OCT) system given a non-Gaussian light source spectrum. By using dispersive materials in the reference arm of the OCT system, the resultant dispersion variation led to a full-width at half maximum (FWHM) of the interference fringe envelope smaller than the Fourier transform-limited value of a Gaussian spectral shape with the same spectral FWHM, at the expense of significant tails. The effects of the tails, which would blur the OCT images, were tremendously reduced with a retrieval algorithm. Simulation results and processed OCT scanning images have shown the capability of the proposed technique.

Journal Article↗

PubMed2002

Evanescent wave coupling in light scattering of an off-axis normally incident Gaussian beam by two parallel nonabsorbing cylinders.

A theoretical and numerical investigation is devised for resonant light scattering of an off-axis normally incident Gaussian beam by two parallel nonabsorbing cylinders based on the related beam theory developed in J. Opt. Soc. Am. A 14, 640 (1997). By varying the half-beam width, we show that the multireflection process between the two scatterers can be minimized. Moreover, the study is an attempt to understand the underlying physics present in the process of resonance excitation by evanescent wave coupling.

Journal Article↗

PubMed2003

Saccadic and perceptual performance in visual search tasks. II. Letter discrimination.

Can the oculomotor system use shape cues to guide search saccades? Observers searched for target letters (D, U, or X) among distractors (the letter O in the discrimination task and blank locations in the detection task) in Gaussian white noise. We measured the accuracy of first saccadic responses on each trial and perceptual (i.e., button-press) responses in separate trials with the stimulus duration chosen so that the saccadic and perceptual processing times were matched. We calculated the relative efficiency of saccadic decisions compared with perceptual decisions, eta(rel) = (d'(sac)/d'(per))2. Relative efficiency was low but consistently greater than zero in discrimination tasks (15% +/- 6%) and high in detection tasks (60% +/- 10%). We conclude that the saccadic targeting system can use shape cues, but less efficiently than the perceptual system can.

Adult↗

PubMed2005

Fiber optic interferometry: statistics of visibility and closure phase.

Interferometric observations with three telescopes or more provide two observables: closure phase information and visibility measurements. When single-mode interferometers are used, both observables have to be redefined in the light of the coupling phenomenon between the incoming wave front and the fiber. We introduce the estimator of both the so-called modal visibility and the modal closure phase. Then we compute the statistics of the two observables in the presence of partial correction by adaptive optics, paying attention to the correlation between the measurements. We find that the correlation coefficients are mostly zero and in any case are never greater than 1/2 for the visibilities and 1/3 for the closure phases. From this theoretical analysis, a data-reduction process using classic least-squares minimization is investigated. In the framework of the AMBER instrument, the three-beam recombiner of the Very Large Telescope Interferometer (VLTI), we simulate the observation of a single Gaussian source and study the performances of the interferometer in terms of diameter measurements. We show that the observation is optimized, i.e., that the signal-to-noise ratio (SNR) of the diameter is maximal when the FWHM of the source is roughly 1/2 of the mean resolution of the interferometer. We finally point out that, in the case of an observation with three telescopes, neglecting the correlation between the measurements leads to overestimating the SNR by a factor of square root of 2. We infer that in any case this value is an upper limit.

Journal Article↗

PubMed1988

Second-trimester maternal serum alpha-fetoprotein levels in pregnancies associated with gastroschisis and omphalocele.

This population-based study analyzes maternal serum alpha-fetoprotein (MSAFP) distributions for 20 cases of gastroschisis and 13 cases of omphalocele occurring in singleton pregnancies from among 72,782 second-trimester pregnancies in Maine and Rhode Island screened consecutively between January 1, 1979 and February 28, 1987. Median values (and ranges) for the two lesions were 4.1 multiples of the median (0.5-29.8) for omphalocele and 7.0 multiples of the median (3.6-13.5) for gastroschisis. The MSAFP distributions for the two conditions were both log-Gaussian, and the log standard deviation was smaller for gastroschisis than for omphalocele. The MSAFP screening sensitivity was greater for gastroschisis than for omphalocele at any given cutoff, and the overall sensitivity of this screening process for detecting open ventral wall defects will differ, therefore, depending upon the relative proportion of gastroschisis and omphalocele cases that occur in the screened population.

Female↗

PubMed2002

Neuronal differentiation and synapse formation occur in space and time with fractal dimension.

The analysis of a set of experimental data obtained by an independent team of researchers confirms that neuronal differentiation or synapse formation do occur in time and space with fractal dimension. The interacting cells create first a dynamic system with its own attractor, (i.e., a fragment of time and space where the dynamic processes occur and where no further evolution of the system is possible at all owing to the action of the intrasystemic forces unless some extrasystemic forces act upon it). This attractor is then modified in the active manner by the differentiating cells until the system attains a degenerated stationary state and differentiation ends. The fractal structure of the system is also lost in the course of tumor progression. Our data indicate that the cellular system can attain the degenerated stationary state, leaving the attractor with a fractal dimension directly or undergoing diversification into many attractors and going through the areas of deterministic chaos. Since evolution of the cellular system is driven by the cooperative dynamic processes, as reflected by the changes of the mean fractal dimension between the intervals of the Gompertzian curve, it is likely that cells differentiate into neurons and create synapses with a conjugated probability and non-Gaussian distribution rather than with the classical probability and the Gaussian distribution. These findings can help to optimize features of artificial neural networks. They also define a simple in vitro biological model for biophysical and biochemical studies on natural neural networks.

Cell Differentiation↗