Using SPSS to do two yuan logistics regression to make predictions

Before we do the logistics regression, we need to make a correlation analysis of the variables you want to predict, and find out the arguments related to your dependent variables. I will not do it here, directly with my data after processing.

Open the data we want to analyze, click "Analyze", select "Regression", and then select "Two Yuan Logistics regression", pop up the following interface,

　　　　

Move the purchase to the dependent variable box, move the consumption amount and consumption amount into the Covariance box, then click the "Save" button, Pop "logistics Regression: Save" screen, select "Predicted value" under "probability", then I love click the browse button, save the model to the table you want to save, Click "Continue" after completion, go back to the interface just after clicking the "OK" button, the "Logistics regression analysis".

It will add a new column of data in your original data table, this is the probability value of that event, in the two logistics regression, the results are expressed in probability values, but in 0 to 0.5 is not the case, 0.5 to 1 means that occurs.

The results of binary logistics regression analysis The most important one is the following table: variables in equations

　　　　

The second column in the table is the coefficient of the regression equation, written as a regression equation:

Logit (P) = 0.01* consumption + (-2.725) * Amount of consumption

"Constant" because the significance of 0.881 is greater than 0.05, so it can be said that the impact of small, can be ignored, add in can also (that constant I asked someone else, said can be ignored, but not sure, afraid in case you add in the test, compared to the results).

The next step is to make predictions using established models.

Open the data you want to predict, then click Utilities, select the Scoring wizard, browse to the address of the model you just saved, and there are a few points to note, looking directly at the graph:

　　　　

　　　　

The next step is to go ahead and then just complete it and add a new column of data to the data table you want to predict, which is the prediction result.

　　　　

So our two-dollar logistics regression forecast will be over.

Attention:

1, here I just carried out a simple two-yuan logistics regression analysis, that is, the dependent variable only two: Yes and no, occurs or does not occur, in fact, there are a variety of circumstances, such as: High, low and three kinds of situations.

2, there is also no processing of variables here, if your data has a lot of variables you must be the first to reduce the dimension, I am here just based on some of my previous data analysis, not specifically to do those preparatory work.

　　　　

　　　　

Using SPSS to do two yuan logistics regression to make predictions