Logistic regression model and its application

To improve the forecasting accuracy of the multinomial qualitative dependent variable by using logistic model,ternary logistic model is established for actual statistical data based on binary logistic regression model.The significance of independent variables is tested by using the likelihood ratio test method to remove the non-significant variable.A linear regression function is determined for each category dependent variable,and the models are tested.The analysis results show that logistic regression model has good predictive accuracy and practical promotional value in handling regression analysis of qualitative dependent variable.