Data-analysis and Inverse Methods in Astronomy, spring 2012

Course page for Data-analysis and Inverse Methods in Astronomy - Tähtitieteen data-analyysi ja inversiomenetelmät

Advanced Course, 7 credits, 53834, Spring 2012, Periods 3 and 4

Course is held in Physicum, Wednesdays at 14-16 in D116 and Fridays at 12-14 in D115.

Lecturers: Mika Juvela, Karri Muinonen, Heikki Haario and Antti Penttilä. Further information.

Lectures and schedule

• 18.1. Introduction. Probability distributions, statististical tests, and error estimation. Luentokalvot. (Mika Juvela)
• 20.1. First steps in the analysis of new data. Linear models. Slides (Mika Juvela).
• 25.1. Generalised linear models and general model fitting. Slides (Mika Juvela).
• 27.1. Principal component analysis and related methods. Slides (Mika Juvela).
• 1.2. Introduction to statistical inference, including Bayesian methods Slides (Antti Penttilä).
• 3.2. Markov chain Monte Carlo. I. Theory MCMC Literature MCMC Theory (Karri Muinonen).
• 8.2. MCMC in practice – tools, convergence and results. Slides (Antti Penttilä).
• 10.2. NOTE: at 12-16 in D211, MCMC. Slides (Heikki Haario, 4 hours).
• 15.2. NOTE: Canceled and rescheduled to Fri 2.3.
• 17.2. NOTE: at 9-13 in D211, MCMC continued.(Heikki Haario, 4 hours).
• 22.2. NOTE: in D211, Gibbs sampling. Slides and data file for exercise (Antti Penttilä).
• 29.2. Visualisation, descriptive methods on multivariate data, image deconvolution (Mika Juvela).
  • Pdf-file on image deconvolution.
  • Links to visualisation tools discussed on the lectures: Ggobi, MayaVi
• 2.3. at 12-14 in Exactum B119: MCMC with virtual observations (Karri Muinonen)

The new deadline for the exercises and small projects is May 2, 2012. We can get together to assess the exercises and projects on May 3, 14-16 at D115. The home exam will take place within May 4-25, 2012. Large projects are due on May 25.

Exercises

Complete and return exercises and small before 2.5.

• Exercises
  • In Exercise 9, calculate the uncertainties at the wavelength of 100µm.
  • Examples of non-linear least squares and optimisation with Matlab/Octave/Python can be found here
• Small projects

Literature

• Advanced Theory of Statistics: Bayesian Inference v. 2B (Kendall's advanced statistics library), A. O'Hagan, M. Kendall, A. Stuart (Author)
• General statistics
  • Mike Irwin (CASU) lectures.
  • Lennart Lindegren (Lund) Statistics for Astronomers.
  • Summer School in Statistics for Astronomers V.
  • L. Lindegren, Lund Observatory: Statistical and Numerical Tools in Astrophysics I.
  • B. Nikolic, Cambridge University: Statistics for astronomy.

Data files

• outliers.txt
• nonlinear.txt
• spectra.dat
• gibbs data
• obs.dat
• dataRed.dat
• dataBlue.dat