Shulou(Shulou.com)06/01 Report--

R language based on Logistic bank loan delinquency rate user analysis is what, I believe that many inexperienced people do not know what to do, so this paper summarizes the causes of the problem and solutions, through this article I hope you can solve this problem.

Today, logistic regression, one of the generalized linear regression models, participated in the school creation project last year. The topic of our group is' Research on female Travel Safety based on logistic Didi Taxi'. Therefore, we have some understanding of this model. Logistic regression is mostly used in medical statistics, and the dependent variables are qualitative variables, which can be ordered and graded, such as sick, non-sick, satisfied, general, dissatisfied, and so on. Logistic regression is often used to find risk factors, such as what are the risk factors of a disease? What is the probability of predicting the occurrence of a disease? To judge how likely someone is to belong to a disease.

1. Data preparation

Before modeling, there must be some data support.

Intercept some of the data as follows

two。 Model theory preparation

Logistic regression belongs to probabilistic nonlinear regression, which is divided into two-classification and multi-classification regression models. For binary Logistic regression, the dependent variable y has only two values of "yes" and "no", marked as 1 and 0. Suppose that the probability of y taking "yes" is p, then the probability of "no" is 1murp, under the action of the independent variable x1magentin x2rem. The purpose of this paper is to study the relationship between the probability p of occurrence when y takes "yes" and the independent variable x1Mague x2pr.

When there is multicollinearity between independent variables, the regression coefficient estimated by least square estimation will be inaccurate, and the improved estimation methods to eliminate multicollinearity mainly include ridge regression and principal component regression.

3. Modeling preparation

Modeling steps of Logistic regression Model

1) set index variables (dependent variables and independent variables) according to the purpose of analysis, and then collect data.

2) the probability of y taking 1 is paired P (yearly 1 | x), and the probability of taking 0 is 1 Murp. The linear regression equation is listed by Ln and independent variables, and the regression coefficient in the model is estimated.

3) carry on the model test: according to the F value and p value in the output analysis of variance table to test whether the regression equation is significant, if the p value is less than the significant level a, the model can pass the test, and the next regression coefficient test can be carried out; otherwise, the index variables should be re-selected and the regression equation should be re-established.

4) carry on the significance test of the regression coefficient: in the multiple linear regression, the significance of the regression equation does not mean that each independent variable has a significant influence on y. In order to eliminate those secondary and dispensable variables from the regression equation, to re-establish a more simple and effective regression equation, it is necessary to test the significance of each independent variable, and the test results are obtained from the parameter estimation table. The stepwise regression method is used to eliminate the least significant dependent variables and reconstruct the regression equation until the model and the participating regression coefficients pass the test.

5) Model application: input the value of independent variable, you can get the value of predictive variable, or control the value of independent variable according to the value of predictive variable.

4. come to conclusion

Logistic regression model program

# read in data

Data

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.