Heliophysics Science Division
Sciences and Exploration Directorate - NASA's Goddard Space Flight Center

July 8, 2011, 12:00 pm - 1:00 pm

July 8, 2011, 12:00 pm - 1:00 pm

Using Bayesian Probability Theory to Solve Large-Dimension Minimization Problems



Dr. Tyson Littenberg (NASA GSFC/University of Maryland)

Bayesian probability theory is becoming an increasingly common foundation for data analysis. Many analysis problems can be recast in a multidimensional optimization/minimization light for which algorithms built from Bayes' theorem, the Markov Chain Monte Carlo (MCMC) family included, are particularly well suited. I will introduce the MCMC approach and demonstrate its capacity to be used as an ``end to end'' data analysis pipeline -- finding the ``best fit'' parameters, characterizing the measurment error, and assigning confidence to the model being used to describe the data. Warning: I am not a heliophysicist, so examples will be in the context of Gravitational Wave astronomy. However, the discussion will be kept as generic as possible.