Po částech lineární regrese

Abstract

This work is about segmented linear regression. It contains a brief introduction to Bayesian statistics, then we define a model for piecewise linear regression. Posterior distribution for the model doesn't have a closed form, so we use RJMCMC algorithm to aproximate the distribution. Before introducing RJMCMC we give an overview of classical MCMC algorithms. At the end of the thesis we show examples of how RJMCMC works. We implemented the RJMCMC algorithm as part of the thesis, for that we used the programming language Python.