Material Detail

Fully Bayesian Source Separation with Application to the CMB

Fully Bayesian Source Separation with Application to the CMB

This video was recorded at MUSCLE Conference joint with VITALAS Conference. Blind source separation refers to the inferring of the values of variables (known as sources) from observations that are linear combinations of them. The observations and sources are usually vectors. Both the sources and the matrix of linear coefficients may be unknown. Here we describe an approach where the sources are assumed to be Gaussian mixtures. An MCMC procedure has been developed that computes the posterior distribution of sources and the matrix of linear coefficients from observations. It is applied to source separation in multi-channel extra-terrestrial microwave data, with the goal of separating out the cosmic microwave background signal.
Rate

Quality

  • User Rating
  • Comments
  • Learning Exercises
  • Bookmark Collections
  • Course ePortfolios
  • Accessibility Info

More about this material

Comments

Log in to participate in the discussions or sign up if you are not already a MERLOT member.
hidden