Mixture regression models

In this thesis, mixture regression concept which is recently defined by combining with advantages of regression analysis and mixture distribution methods, has been included. With datas taken from homogeneous structured population, generally biased estimations are obtained. In these conditions, the applications of mixture regression models, also known as latent class regression analysis, have been advocated. After the examinations of performed illustrations associated with mixture regression models and analysis, the strengths and limitations of this method are discussed in the studies. Using AMOS sowftware program, mixture regression model, analysis, distributions and hypothesis testing have been mentioned on an artificial dataset. Finally, AMOS have been discussed about addition on mixture regression analysis and obtained results compared with alternative programs.Keywords: Mixture Distribution, Linear Mixture Regression Models, Latent ClassRegression, Mixture Regression Analysis, Bayesian Estimation.