Conor Lawless

12 August 2019

Inference for dynamic simulation models

I recently bought a copy of the latest edition of Darren Wilkinson’s book: Stochastic Modelling for Systems Biology because I’m interested to see what he has to say about Approximate Bayesian Computation (ABC). Browsing through, I am reminded of the very useful example in Chapter 11: Inference for stochastic kinetic Models, which describes a likelihood-free method for Bayesian inference of dynamic simulation model parameters. This approach involves using a bootstrap particle filter for marginal likelihood estimation. ... Read more

21 November 2017

Visualising correlation in an AR(1) process

Autoregression models are a type of stochastic, dynamic process. They are a mathematical representation of some value that varies with time, where the variation includes a random, unpredictable component. Using a computer to generate (pseudo-)random numbers, we can generate a set of simulated values across time that are consistent with this kind of model. Stochastic simulations aim to capture the random component in the process and so are usually different every time they’re run. ... Read more