Search results for "sparse"

Numerical Study of Two Sparse AMG-methods

2003

A sparse algebraic multigrid method is studied as a cheap and accurate way to compute approximations of Schur complements of matrices arising from the discretization of some symmetric and positive definite partial differential operators. The construction of such a multigrid is discussed and numerical experiments are used to verify the properties of the method.

A Douglas–Rachford method for sparse extreme learning machine

2019

Phase retrieval of a Kolmogorov phase screen from very sparse data using four binary masks

2020

We investigate experimentally the phase retrieval of a Kolmogorov phase screen from very sparse data by modulating its amplitude with four binary masks and compare the retrieved phase screen to the ground truth measured with a surface profiler. Previously, we have shown in simulations that this kind of modulation can be successfully used for the phase retrieval of a Kolmogorov phase screen. After subtracting the ground truth from the retrieved phase screen, the root-mean-square error decreased from 0.14 µm to 0.10 µm. We conclude that a Kolmogorov phase screen can be recovered using simple modulation and very sparse data.

Method specific Cholesky decomposition : Coulomb and exchange energies

2008

We present a novel approach to the calculation of the Coulomb and exchange contributions to the total electronic energy in self consistent field and density functional theory. The numerical procedure is based on the Cholesky decomposition and involves decomposition of specific Hadamard product matrices that enter the energy expression. In this way, we determine an auxiliary basis and obtain a dramatic reduction in size as compared to the resolution of identity (RI) method. Although the auxiliary basis is determined from the energy expression, we have complete control of the errors in the gradient or Fock matrix. Another important advantage of this method specific Cholesky decomposition is t…

DgCox: a differential geometric approach for high-dimensional Cox proportional hazard models

2014

Many clinical and epidemiological studies rely on survival modelling to detect clinically relevant factors that affect various event histories. With the introduction of high-throughput technologies in the clinical and even large-scale epidemiological studies, the need for inference tools that are able to deal with fat data-structures, i.e., relatively small number of observations compared to the number of features, is becoming more prominent. This paper will introduce a principled sparse inference methodology for proportional hazards modelling, based on differential geometrical analyses of the high-dimensional likelihood surface.

Preliminary development of a questionnaire to measure the extra-pulmonary symptoms of severe asthma

2021

Abstract Background Research into the effects of asthma treatments on the extra-pulmonary symptoms of severe asthma is limited by the absence of a suitable questionnaire. The aim was to create a questionnaire suitable for intervention studies by selecting symptoms that are statistically associated with asthma pathology and therefore may improve when pathology is reduced. Methods Patients attending a specialist asthma clinic completed the 65-item General Symptom Questionnaire (GSQ-65), a questionnaire validated for assessing symptoms of people with multiple medically unexplained symptoms. Lung function (FEV1%) and cumulative oral corticosteroids (OCS) calculated from maintenance dose plus ex…

Feature selection on a dataset of protein families: from exploratory data analysis to statistical variable importance

2016

Proteins are characterized by several typologies of features (structural, geometrical, energy). Most of these features are expected to be similar within a protein family. We are interested to detect which features can identify proteins that belong to a family, as well as to define the boundaries among families. Some features are redundant: they could generate noise in identifying which variables are essential as a fingerprint and, consequently, if they are related or not to a function of a protein family. We defined an original approach to analyze protein features for defining their relationships and peculiarities within protein families. A multistep approach has been mainly performed in R …

Low-Rank Tucker-2 Model for Multi-Subject fMRI Data Decomposition with Spatial Sparsity Constraint

2022

Tucker decomposition can provide an intuitive summary to understand brain function by decomposing multi-subject fMRI data into a core tensor and multiple factor matrices, and was mostly used to extract functional connectivity patterns across time/subjects using orthogonality constraints. However, these algorithms are unsuitable for extracting common spatial and temporal patterns across subjects due to distinct characteristics such as high-level noise. Motivated by a successful application of Tucker decomposition to image denoising and the intrinsic sparsity of spatial activations in fMRI, we propose a low-rank Tucker-2 model with spatial sparsity constraint to analyze multi-subject fMRI dat…

Domestication of a Robotic Medication-Dispensing Service Among Older People in Finnish Lapland

2020

This paper presents a case study on a robotic medication-dispensing service used in the everyday lives of older people in Finnish Lapland through the concept of domestication. The study took an ethnographic approach. A total of 11 service users, practical nurses, and other health-care professionals participated; the service users averaged age 81 years (M= 81.4, SD = 5.4). The data comprised semistructured interviews complemented by observations and photographs at service users’ homes. We concluded that the domestication of the service was successful, although the service users sometimes felt that it limited their lives. The service users stated that learning and subsequently using the servi…

A probabilistic compressive sensing framework with applications to ultrasound signal processing

2019

Abstract The field of Compressive Sensing (CS) has provided algorithms to reconstruct signals from a much lower number of measurements than specified by the Nyquist-Shannon theorem. There are two fundamental concepts underpinning the field of CS. The first is the use of random transformations to project high-dimensional measurements onto a much lower-dimensional domain. The second is the use of sparse regression to reconstruct the original signal. This assumes that a sparse representation exists for this signal in some known domain, manifested by a dictionary. The original formulation for CS specifies the use of an l 1 penalised regression method, the Lasso. Whilst this has worked well in l…