Multinomická regrese pro nominální a ordinální vysvětlovanou proměnnou

AbstractThe diploma thesis deals with the method of multinomial regression with nominal and ordinal response variables. The thesis is divided to three different parts. The first part deals with the basic procedures of linear regression. Briefly, the basic procedures of single and multiple linear regression are descibed. The name of the second part is logistic regression. At the beginning of this section there is a logistic regression with a binary response variable. Logistic regression with a binary response variable occurs at thesis because i tis closely related to multinomial regression. The second part of this chapter is detailed description of multinomial regression with nominal response variable. There are methods for estimating parameters, testing estimated parameters, and predicting the new observations. The last part of the second chapter is a detailed description of multinomial regression methods with ordinal response variable. The third chapter deals with the practical use of multinomial regression. It is divided into three blocks. The first block is a description of the dataset EUROSTUDENT VI. The second block deals with the estimation of the regression model with the nominal response variable. The last block deals with the estimation of the regression model with the ordinal response variable.