Attention should be paid to how much correlation between independent variables is in regression anal

This is good for multiple linear regression, but it is actually binary linear regression, with 2 independent variables A and B, and the dependent variable C.

The univariate linear regression equation y=ax+b, the coefficient a 0, y is positively correlated with x, when x is high, y is high, when x is low, y is low, and a 0 is opposite.

The binary linear regression equation is y=ax1+bx2+c,x1,x2 corresponding to the a and b variables of this problem.

If the coefficients a and b are both positive, then when a is high and b is high, c will also be high.

If the coefficient is negative, then C will be low when A is high and B is high.

If the coefficient A is positive and B is negative, then A is high, B is low, and C will be high, but A is low and B is high, and the effect is subtracted, and it is difficult to determine the level of C.

The same goes for a negative case and B is positive.

Operation steps: analysis-regression-linear, c is the dependent variable, a, b is the independent variable, if the p value of the anova table is less than the regression method is established, you can follow the above steps.

If it is greater than that, it means that the linear model is not valid, then the nonlinear model needs to be considered for correlation analysis, and the reason is the same.

Review:Regression analysisandDependent variableThe nature of any system (or model) is that any system (or model) is made up of various variables, and when we analyze these systems (or models), we can choose to study the effects of some of these variables on others.

Then the variables we choose are called independent variables, and the amount that is affected is called the dependent variable.

The modern interpretation of the term regression is very succinct, a statistical analysis method that studies the dependence of dependent variables on independent variables.

The aim is to estimate or play the mean of the variable by the given value of the independent variable. It can be used for time series modeling, and for discovering causal relationships between various variables.

There are many benefits to using regression analysis, such as indicating a significant relationship between the independent and dependent variables, or indicating the strength of the effect of multiple independent variables on a dependent variable.

Regression analysis can also be used to compare the interaction between variables measured by different measures, such as the association between change and the number of activities.

These benefits benefit market researchers, data grinding analysis.

People and data scientists exclude and measure the best set of variables to build a model.

Summary. It is not necessary to continue regression with only one variable, and if it is necessary to continue regression in the case of multiple variables.

Correlation and regression are pretty much the same thing when there are only two variables. In the case of multivariate, you can use regression to do **, consider moderating variables, collinearity problems, and multiple regression and some other functions, so continue to do regression, or two variables, it is really not necessary, if it is multivariate, it can still be considered.

Because Pearson correlation analysis is a simple, general representation of the correlation between variables, it does not consider whether there is collinearity or mutual influence between variables. Therefore, when other correlation analyses can be done, such as regression analysis, analysis of variance, etc., there is no need to look at the results of Pearson correlation analysis, but to use the data of regression analysis as the basis.

If the correlation analysis has only one independent variable, do you still need to perform regression analysis?

Only one variable is not required to continue regression, and if it is multivariate, it is necessary to continue regression. Correlation and regression are pretty much the same thing when there are only two variables. In the case of multivariate, you can use regression to do **, consider the moderating variables, collinearity problems, and multiple regression and some other functions, so, continue to do regression, Hongqiao is still two variables, it is really not necessary, if it is multivariate, it can still be considered.

Because Pearson correlation analysis is a simple, general representation of the correlation between variables, it does not consider whether there is collinearity or interaction between variables. Therefore, when other correlation analyses can be done, such as regression analysis, analysis of variance, etc., there is no need to look at the results of Pearson correlation analysis, but to use the data of regression analysis as the basis.

I see. Uh-huh.

What is the ROC curve to test or something, and can ROC be done between two variables?

If you don't understand anything, you can continue to ask me.

Generally speaking, there can be a group comparison method for the two diagnostic methods, the group comparison method is the two diagnostic methods for different patients who are tested by the hall, and the paired comparison method is for the same subject to receive two different diagnostic methods. The ROC curve is suitable for variables that reflect the effect or outcome of the categorical group.

A ROC curve can be made between two variables.

The regression control variables are three suitable. The selection of the number of control variables in the regression model is mainly based on economic theory, generally speaking, the number of three control variables is too small, and there may be the problem of missing variables, which may lead to unreliable regression results.

Introduction to variablesVariables** are abstract concepts in computer languages that can store the results of calculations or represent values. Variables can be accessed by variable name. In imperative languages, variables are usually mutable.

But in a purely functional language, variables can be immutable. In some languages, variables may be explicitly abstractions that represent mutable states with storage space.

Other languages, however, may use other concepts to refer to this abstraction, rather than strictly defining the exact extension of the variable. Variables are useful because they allow you to give a short, easy-to-remember name to every piece of data that you are going to use in your program. Variables are useful because they allow you to give a short, easy-to-remember name to every piece of data that you are going to use in your program.

Summary. Hello, you can use multivariate ANOVA instead of linear regression analysis to do it, through multivariate ANOVA, you can analyze multiple dependent variables and multiple independent variables at the same time, and then you can also do parameter estimation to get regression coefficients and fitting values.

Hello, you can use multivariate ANOVA instead of linear regression analysis to do it, through multivariate ANOVA, you can analyze multiple dependent variables and multiple independent variables at the same time, and then you can also do parameter estimation to get regression coefficients and fitting values.

The above information is for reference. The time to answer the question is always short, I have finished it for you, I hope you are satisfied! Hope to get your likes!

How to do it.

You can use multivariate ANOVA instead of linear regression analysis, and with multivariate ANOVA, you can analyze multiple dependent variables and multiple independent variables at the same time, and then you can also do parameter estimation to get regression coefficients and fit values.

In the logistics regression, twoArgument.

Dependent variableand independent variables into a list of grids, with the dependent variable at the top and the independent variable at the bottom (single variable pulls in one, multi-factor pulls in multiple).

To set up the regression method, choose the simplest method: enter, which refers to incorporating all the variables into the equation at once. The methods are all step-by-step approaches.

Hierarchical data and continuous data do not need to set dummy variables. Multi-categorical variables need to be set up with four categories of dummy variables, ABCD, with A as a reference, B has no effect relative to A, C has no effect relative to A, and D has no effect relative to A.

Principle:

If it is directly regressed linearly.

The model is deducted into the logistic regression, which will cause different bilateral value intervals and a general non-linear relationship between the two sides of the equation. Because the dependent variable in logistic is a dichotomous variable, the estimated value of a probability as the dependent variable of the equation is 0-1, but the range of the value on the right side of the equation is infinity or infinitesimal size.

That's why logistic regression is introduced.

Hello, I'm glad to serve you and give you the following answer: Not necessarily, but it may be important for control variables to have a blind impact on the outcome. When the control variable is not significant, it may cause bias in the model, which is the so-called "control variable missing" problem, which may affect the accuracy and confidence of the regression results.

Increasing the number of control variables can improve the accuracy of the model, and thus better control the significance of the variables. 3.Try different models:

Experimenting with different models, such as analytic hierarchy process, time series analysis, multiple linear regression, etc., can give you better control over the significance of variables. 4.Try different evaluation metrics:

Trying different evaluation indicators, such as F-test, R2, etc., can better control the significant pattiness of the variables. 5.Transform the control variable:

Transforming control variables, such as logarithmic transformation, squared transformation, etc., can better control the significance of the variables.

Correlation and regression analysis are required to study whether there is a linear relationship between two variables that are professionally related and how to obtain a linear regression equation. From the perspective of the purpose of the study, if it is only to understand the closeness and direction of the linear relationship between two variables, it is advisable to use linear correlation analysis. If the linear regression equation is only to establish a linear regression equation for calculating the dependent variable from the independent variable, it is advisable to use linear regression analysis.

In terms of the conditions of the data, the correlation analysis requires that both variables are random variables (e.g., human length and weight, blood selenium and hair selenium); Regression analysis requires that the dependent variable be a random variable, and the independent variable can be random or general (i.e., the value of the variable can be specified in advance, such as the dose of medication).

The remaining 31%.