ABSTRACT
Logistic regression model is a type of statistic analytical method, which is usually used by dichotomous variables. Different from linear regression, Logistic regression is a non-linear model. Parameter estimation is Maximum likelihood estimate method. Logistic regression method can set up models for categorical independent variable, categorical dependent variable, and there is a whole set of criterion to test regression models and the parameters. The regression model may be calculated with Ordinarily Least Square. At first, individual data is grouped according to independent variable. Secondly, every group event probability is estimated. Then change every group event probability into logarithm odds which see as dependent variable of linear regression model. Combined with Geographical System Information technology and statistics technology, regional slope stability has been evaluated. The probability of the first model is regarded as the Bayesian statistic inference’s priori information.
