Skip to main content
Fig. 1 | BMC Medical Research Methodology

Fig. 1

From: High performance implementation of the hierarchical likelihood for generalized linear mixed models: an application to estimate the potassium reference range in massive electronic health records datasets

Fig. 1

Performance of GLMM fitting methods in the simulated Poisson datasets: fixed effects (A), variance components (B) and random effects (C). For set of parameters we show the standardized Bias (stdBias) and the Mean Square Error (MSE). Abbreviations: AGH0(1): Adaptive Gaussian Quadrature of order 0 or 1, glmmTMB: estimates returned by the relevant package in R, h-lik: the proposed method in text, iGLM: interconnected Generalized Linear Model, MCMC: Markov Chain Monte Carlo (MCMC) posterior mean, MCMCmode: mode of posterior marginals from MCMC simulations. N of IP: Number of Individual Patients in each simulated dataset

Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us