Logistic regression, when is a "True Positive"?

23 Views Asked by At

I have the following table, (LR = Logistic Regression; BA = Biological activity):

LR  BA
0.539   0
0.526   0
0.527   0
0.597   0
0.508   0
0.536   0
0.507   0
0.546   0
0.552   0
0.456   0
0.529   0
0.527   0
0.454   0
0.546   0
0.528   0
0.531   0
0.762   1
0.527   0
0.721   1
0.517   0
0.531   0
0.618   1
0.490   0
0.501   0
0.483   0
0.571   0
0.491   0
0.496   0
0.464   0
0.488   0
0.507   0
0.523   0
0.491   0
0.487   0
0.490   0
0.489   0
0.502   0
0.702   1
0.479   0
0.505   0
0.488   0
0.496   0
0.489   0
0.546   1
0.492   0
0.578   1
0.548   0
0.526   0
0.491   0
0.531   0
0.491   0
0.511   0
0.494   0
0.507   0
0.533   0
0.729   1
0.707   1

Looking at the LR values:

  • Would be a True Positive those values larger than 0.5?

Looking at the LR values and the BA:

  • The LR frontier value to define the True Positive would be the minor value that matches the BA positive value?