Robustification of regression model with the fixed and random effects
|Author:||Bc. Magdaléna Raušová|
|Year:||2014 - summer|
|Leaders:|| prof. RNDr. Jan Ámos Víšek CSc.
|Work type:|| Bachelors
|Awards and prizes:||B.A. with distinction from the Dean of the Faculty of Social Sciences for an extraordinarily good bachelors diploma thesis.|
|Abstract:||In case of some influential observations in an econometric analysis, the classical methods,
such as ordinary least squares, are likely to fail. The problem of outliers and leverage
points can be overcome by the robust methods. This thesis studies the use of robust
methods for panel data - specifically, the robustified versions of the methods of fixed
and random effects utilizing the least weighed squares are studied. After introducing
the theoretical background, results of a numerical study are provided. This numerical
study is a Monte Carlo study that shows, how the classical and robust methods work
under several levels of contamination and also, how the choice of the weight function can
influence the results of the methods that utilize the least weighted squares.