如何解释逻辑回归模型中的分类表

Question

This question was migrated from Stack Overflow because it can be answered on Cross Validated. Migrated 4 days ago .

I completed a logistic regression model and a classification table but I am unsure of how to interpret the results of this table.

The output is as follows

                   Predicted_Value
       Actual_Value FALSE TRUE
                 0   515   37
                 1    89  109

Answer 1

Classic logistic regression outputs a probability [0-1] for a patient to have the value "1" of your actual observed values when you train the model.

In order to get a binary Predicted value, then you need to put a threshold on your outputed vector of probabilities.

This is what has been done in order to get the contingency table assessing how accurate your prediction model is when compared with the actual value.

Here, you can compute for example Accuracy, Sensitivity, Specificity.

Accuracy = (109 + 515) / sum(tab) = 83.2% correctly predicted patients  
Sensitivity = 109 / (109 + 89) = 55.0% correctly predicted Positive patients  
Specificity = 515 / (515 + 37) = 92.3% correctly predicted Negative patients  

If you change your cut-off then you will have more or less positively predicted patients ; this will impact your performance criteria so the choice of cut-off value is yours.

Here we don't talk about training and validation sets but these informations are important if you want to know whether your model is robust or not.