Presentation
Approximate bayesian computation (ABC) methods are standard tools for inferring parameters of complex models when related likelihood functions are analytically intractable. A common approach to accelerate the acceptance rate of the rejection sampling based ABC methods is to use sequential monte carlo (ABC SMC) to produce an adapting sequence of proposal distributions. Proposal distributions for the next iteration round are most often calculated empirically from the weighted set of samples of the current iteration round. We investigate methods that improve the acceptance rate by taking the simulated observation values into account in construction of the proposal distribution and introduce a method called stratified distance ABC SMC. Algorithm stratifies samples based on their distance between corresponding simulated and the true observation and estimates a distinct proposal distribution for each set of samples. Further efficiency is gained by introducing a novel stopping rule for the sequential process based on the stratified posterior samples.
Program
11:30 – Doors open and lunch is served
12:00 – "Accelerating ABC SMC" by Henri Pesonen (Researcher, OCBE, Institute of Basic Medical Sciences)
This event is open for all PhD candidates and postdocs. No registration needed.
About the seminar series
Once a month, dScience will invite you to join us for lunch, soft drinks and professional talks at the Science Library. In addition to these, we will serve lunch to PhD candidates in our lounge in Kristine Bonnevies hus every Thursday. Due to limited space (40 people), this will be first come, first served. See how to find us here (download).
Our lounge can also be booked by PhDs and Postdocs on a regular basis, whether it is for a meeting or just to hang out – we have fresh coffee all day long! Read more about the seminar series here.