When: Thursday 16th August 2018, 13:00-14:00.
Included in meeting:
The fortnightly BRAG meeting will be held tomorrow, August 16, 2018, Thursday at 1 pm in Y801.
This week we will have presentations by Anthony and Imke.
Title: Metrics on probability measures: Applications for dynamic queueing networks
Abstract: We show how metrics on probability measures such as maximum mean discrepancy (MMD) and the Wasserstein distance can be used with functional datasets arising from a dynamic queueing network to estimate parameters using a simulated annealing approximate Bayesian computation (SABC) sampler. A key challenge confronting DQN parameter estimation is simulation speed. We use the R package queuecomputer to make simulation-based inference on large DQNs feasible. This is the first example of likelihood-free parameter inference for a DQN. There are applications in hospitals, web-servers and transit networks. We demonstrate the approach using real data from an international airport passenger terminal.
Title: Particle methods for stochastic differential equation mixed effects models
Abstract: Stochastic differential equation mixed effects models (SDEMEMs) are particularly useful for biomedical and pharmacokinetic/pharmacodynamic modelling as they are able to account for both system noise and inter-subject variability. Parameter inference for these models can be difficult however. The transition density of SDEMEMs are rarely available in closed form, which results in an intractable likelihood function. Research on Bayesian methods have largely been focussed on using closed form approximations of the transition density within Markov chain Monte Carlo (MCMC) algorithms. I will present a few applications of particle MCMC methods to SDEMEMs. These methods are useful in the sense that they (generally) only require that the transition density can be sampled from.
We look forward to seeing you all there!
Trish and Farzana