Abstract:
In this project, microlensing events are modelled and analysed using a new methodology employing the MultiNest algorithm. MultiNest is based on the principles of Bayesian inference, which allows us to solve the model selection and parameter estimation problems simultaneously. The focus is placed on the model selection problem since a Bayesian based algorithm such as MultiNest allows us to shift the approach to model selection from qualitative arguments to a quantitative quality factor. The methodology is demonstrated by testing a finite-source point-lens model versus a finite-source binary-lens model as well as testing for the presence of parallax effects. This is done for a simulated synthetic event as proof of concept and for a real event, OGLE-2011-BLG-0251. Nested Sampling and its variant algorithms such as MultiNest have been tried and tested in many fields of study. By demonstrating MultiNest on a real microlensing event, the aim of this project is to provide an impetus for said algorithms to find their place in the microlensing community as well.