Expanse Of Box-Cox Logit Model And Its Improved Form

The main purpose of this thesis is to expand the use of Box-Cox Logit model into the Nested Logit model. And at the same time, we bring forward a new model form which has some better properties than BCL model.As we can see from our paper, deterministic part of the utility is one of the most important factors that can affect the choice probability of logit model, but most of the study in discrete choice model doesn't pay much attention to it. In allusion to this ,we expand the use of Box-Cox Logit model into the Nested Logit model and bring forward an improved form which is more flexible than the BCL model.The study of this paper is based on two foundations. One is the discrete choice model theory. And the other is Box-Cox data conversion theory in statistics research. In order to relax the restriction of IIA property and the linear hypothesis of the deterministic part of the utility, we focus on extending the Multinomial Logit to Nested Logit and on mending the deformation of deterministic part of the utility, and at last we give a demonstration.