arXiv subjects

Danielle Fitt

Publications and source records attributed to Danielle Fitt.

4 recordsLinked to original sources

arXiv2024

Equitable Length of Stay Prediction for Patients with Learning Disabilities and Multiple Long-term Conditions Using Machine Learning

People with learning disabilities have a higher mortality rate and premature deaths compared to the general public, as reported in published research in the UK and other countries. This study analyses hospitalisations of 9,618 patients identified with learning disabilities and long-term conditions for the population of Wales using electronic health record (EHR) data sources from the SAIL Databank. We describe the demographic characteristics, prevalence of long-term conditions, medication history, hospital visits, and lifestyle history for our study cohort, and apply machine learning models to predict the length of hospital stays for this cohort. The random forest (RF) model achieved an Area Under the Curve (AUC) of 0.759 (males) and 0.756 (females), a false negative rate of 0.224 (males) and 0.229 (females), and a balanced accuracy of 0.690 (males) and 0.689 (females). After examining model performance across ethnic groups, two bias mitigation algorithms (threshold optimization and the reductions algorithm using an exponentiated gradient) were applied to minimise performance discrepancies. The threshold optimizer algorithm outperformed the reductions algorithm, achieving lower ranges in false positive rate and balanced accuracy for the male cohort across the ethnic groups. This study demonstrates the potential of applying machine learning models with effective bias mitigation approaches on EHR data sources to enable equitable prediction of hospital stays by addressing data imbalances across groups.

cs.LG↗

arXiv2019

Likely oscillatory motions of stochastic hyperelastic solids

Stochastic homogeneous hyperelastic solids are characterised by strain-energy densities where the parameters are random variables defined by probability density functions. These models allow for the propagation of uncertainties from input data to output quantities of interest. To investigate the effect of probabilistic parameters on predicted mechanical responses, we study radial oscillations of cylindrical and spherical shells of stochastic incompressible isotropic hyperelastic material, formulated as quasi-equilibrated motions where the system is in equilibrium at every time instant. Additionally, we study finite shear oscillations of a cuboid, which are not quasi-equilibrated. We find that, for hyperelastic bodies of stochastic neo-Hookean or Mooney-Rivlin material, the amplitude and period of the oscillations follow probability distributions that can be characterised. Further, for cylindrical tubes and spherical shells, when an impulse surface traction is applied, there is a parameter interval where the oscillatory and non-oscillatory motions compete, in the sense that both have a chance to occur with a given probability. We refer to the dynamic evolution of these elastic systems, which exhibit inherent uncertainties due to the material properties, as `likely oscillatory motions'.

physics.class-ph↗

arXiv2018

Likely cavitation in stochastic elasticity

We revisit the classic problem of elastic cavitation within the framework of stochastic elasticity. For the deterministic elastic problem, involving homogeneous isotropic incompressible hyperelastic spheres under radially symmetric tension, there is a critical dead-load traction at which cavitation can occur for some materials. In addition to the well-known case of stable cavitation post-bifurcation at the critical dead load, we show the existence of unstable snap cavitation for some isotropic materials satisfying Baker-Ericksen inequalities. For the stochastic problem, we derive the probability distribution of the deformations after bifurcation. In this case, we find that, due to the probabilistic nature of the material parameters, there is always a competition between the stable and unstable states. Therefore, at a critical load, stable or unstable cavitation occurs with a given probability, and there is also a probability that the cavity may form under smaller or greater loads than the expected critical value. We refer to these phenomena as `likely cavitation'. Moreover, we provide examples of homogeneous isotropic incompressible materials exhibiting stable or unstable cavitation together with their stochastic equivalent.

cond-mat.soft↗

arXiv2018

Likely equilibria of stochastic hyperelastic spherical shells and tubes

In large deformations, internally pressurised elastic spherical shells and tubes may undergo a limit-point, or inflation, instability manifested by a rapid transition in which their radii suddenly increase. The possible existence of such an instability depends on the material constitutive model. Here, we revisit this problem in the context of stochastic incompressible hyperelastic materials, and ask the question: what is the probability distribution of stable radially symmetric inflation, such that the internal pressure always increases as the radial stretch increases? For the classic elastic problem, involving isotropic incompressible materials, there is a critical parameter value that strictly separates the cases where inflation instability can occur or not. By contrast, for the stochastic problem, we show that the inherent variability of the probabilistic parameters implies that there is always competition between the two cases. To illustrate this, we draw on published experimental data for rubber, and derive the probability distribution of the corresponding random shear modulus to predict the inflation responses for a spherical shell and a cylindrical tube made of a material characterised by this parameter.

cond-mat.soft↗