A new regression model for data on Z
Diagnostic analysis. Estimation by maximum likelihood. Paired count samples. Skew discrete Laplace distribution. Regression models.
There are several practical situations in which it is of interest to model events associated with discrete-valued variables. Until now, theories that have been constructed and refined to handling observations of this nature have an emphasis on modeling non-negative discrete data. Nevertheless, discrete observations that may assume any value on the set of integers Z = {... -2, -1, 0, 1, 2, ...}can also be found in different contexts. The main objective of this master thesis are to propose a new parameterization for the skew discrete Laplace distribution (KOZUBOWSKI; INUSAH, 2006), in terms of the mean and a dispersion parameter, and then define a new regression model able of modeling observations that assume values on Z based on this distribution. We consider the maximum likelihood estimator for the estimation of unknown model parameters. We propose diagnostic methods to evaluate the goodness of fit. We performed some simulation studies to verify the performance of the proposed estimators, test statistics and residuals. Finally, we apply the model to two real data sets.