Lab 5: Bayesin Inference & Non-linear Models
(github)
Start Sept 21 –- Due Sept 26
Goals
In this lab, students will learn to:
Characterize the mass and orbit of an exoplanet using radial velocity observations
Build a Bayesian model for astronomical observations
Specify a prior and likelihood using a probabilistic programming language
Perform posterior sampling
Qualitatively evaluate the convergence of Markov chains
Use an approximate model to assist with global search over a multi-modal parameter space
Assess the sensitivity of inferential results to the choice of prior & likelihood
Students will strengthen their understanding of:
Assessing models
Evaluating the sensitivity of results to choice of priors and likelihood.
Evaluating the effects of model misspecification
Logistics
Follow this link to create your own private copy of this lab's repository on GitHub.com. See the help on the course website for instructions on getting setup to use ACI, cloning, committing, pushing and submitting your work.