Study Of Binary Skew LOGIT Model And Its Application

Among all the statistic models, the linear model is preferred because it is the simplest, the easiest used and the most efficient one. However, in many situations, the linear relationship between the response variable and the explanatory variable does not come into existence. One idea to solve this problem is retaining the linear assumption by modifying the linear model appropriately. Binary Probit model and binary Logit model are two successful realizations of this idea. Another idea is to make nonlinear regression when the linear hypothesis did not stand up.Binary models are originated from the statistical analysis of biological trials. But with the development of the models, they relaxed themselves from the biological trials and had become the independent generalized linear models. Due to their excellent statistical properties, binary Probit model and binary Logit model have become the two of the most widely used binary models today.In this paper, starting with the invention of the binary models, the developing process of the two binary models and the research about other S curves are introduced briefly. Then, the statistical methods for several main models are summarized. The binary Logistic regression model is extended and the binary skew logit model is defined., its nature is also proved ,from which we know that the binary logit model can be seen as a special case of the binary skew logit model,beacause of its greater flexibility,generally ,we can get higher precision in fitting.then we offer the method of parameter estimation bothrepeated observation values cound be gotten and repeated observation values coundn't be gotten. Finally, with the help of SPSS and MATLAB software,one example about the ralationship between information of loans costomer and repayments on time is given, Results displayed that the binary skew Logistic model has more predominance in fitting than the binary Logit model .