arXiv subjects

James Rafferty

Publications and source records attributed to James Rafferty.

4 recordsLinked to original sources

arXiv2026

Bayesian Profile Regression using Variational Inference to Identify Clusters of Multiple Long-Term Conditions Conditioning on Mortality in Population-Scale Data

Multiple long-term conditions (MLTC) are increasingly observed in clinical practice globally. Clustering methods to group diseases into commonly co-occurring clusters have been of interest for further understanding of how MLTC group together and their associated impact on patient outcomes. However, such approaches require large, often population-scale datasets. Bayesian Profile Regression (BPR) is a statistical model that combines a Dirichlet Process Mixture model with a hierarchical regression model, in order to form clusters of items conditional on covariates and an outcome of interest. We developed a BPR model using full-rank Stochastic Variational Inference (SVI) for application in large-scale data. We assessed it's performance using simulation studies comparing fits using the No-U-turn (NUTS) sampler and full-rank SVI. We then fit a BPR model to find clusters of MLTC in a population-scale data held in the Secure Anonymised Information Linkage (SAIL) databank. We found results from full-rank SVI compared well with results from NUTS in a simulation study, and the improved fitting performance allowed for fitting models in population-scale datasets. There were 1,296,463 individuals in our electronic health record (EHR) cohort. The clustering model was conditioned on age at cohort entry, socioeconomic deprivation and sex with mortality as the outcome. We used the Elixhauser comorbidity index disease definitions, and found there were 33 disease clusters. We found that clusters featuring metastatic cancer and cardiovascular diseases, such as congestive heart failure, were most strongly associated with the probability of mortality. Our findings show that SVI can be a useful and accurate method for fitting Bayesian models, especially when the dataset size would make Monte Carlo methods prohibitively time consuming or impossible.

stat.AP↗

arXiv2011

Holographic Roberge-Weiss Transitions II: Defect Theories and the Sakai-Sugimoto Model

We extend the work of Aarts et al., including an imaginary chemical potential for quark number into the Sakai-Sugimoto model and codimension k defect theories. The phase diagram of these models are a function of three parameters, the temperature, chemical potential and the asymptotic separation of the flavour branes, related to a mass for the quarks in the boundary theories. We compute the phase diagrams and the pressure due to the flavours of the theories as a function of these parameters and show that there are Roberge-Weiss transitions in the high temperature phases, chiral symmetry restored for the Sakai-Sugimoto model and deconfined for the defect models, while at low temperatures there are no Roberge-Weiss transitions. In all the models we consider the transitions between low and high temperature phases are first order, hence the points where they meet the Roberge-Weiss lines are triple points. The pressure for the defect theories scales in the way we expect from dimensional analysis while the Sakai-Sugimoto model exhibits unusual scaling. We show that the models we consider are analytic in \mu^2 when \mu^2 is small.

hep-th↗

arXiv2010

Holographic Roberge-Weiss Transitions

We investigate N=4 SYM coupled to fundamental flavours at nonzero imaginary quark chemical potential in the strong coupling and large N limit, using gauge/gravity duality applied to the D3-D7 system, treating flavours in the probe approximation. The interplay between Z(N) symmetry and the imaginary chemical potential yields a series of first-order Roberge-Weiss transitions. An additional thermal transition separates phases where quarks are bound/unbound into mesons. This results in a set of Roberge-Weiss endpoints: we establish that these are triple points, determine the Roberge-Weiss temperature, give the curvature of the phase boundaries and confirm that the theory is analytic in mu^2 when mu^2~0.

hep-th↗

arXiv2009

Real time response on dS_3: the Topological AdS Black Hole and the Bubble

We study real time correlators in strongly coupled N=4 supersymmetric Yang-Mills theory on dS_3 x S^1, with antiperiodic boundary conditions for fermions on the circle. When the circle radius is larger than a critical value, the dual geometry is the so-called "topological AdS_5 black hole". Applying the Son- Starinets recipe in this background we compute retarded glueball propagators which exhibit an infinite set of poles yielding the quasinormal frequencies of the topological black hole. The imaginary parts of the propagators exhibit thermal effects associated with the Gibbons-Hawking temperature due to the cosmological horizon of the de Sitter boundary. We also obtain R-current correlators and find that after accounting for a small subtlety, the Son-Starinets prescription yields the retarded Green's functions. The correlators do not display diffusive behaviour at late times. Below the critical value of the circle radius, the topological black hole decays to the AdS_5 "bubble of nothing". Using a high frequency WKB approximation, we show that glueball correlators in this phase exhibit poles on the real axis. The tunnelling from the black hole to the bubble is interpreted as a hadronization transition.

hep-th↗