Figure 2
The x 1 -x 2 scatter plot and logistic regression boundary. This figure shows the two risk factor scatter plot for cases (diamonds) and controls (multiplication signs). Each point represents a given sample's (x1, x2) risk vector plotted in component form. The solid line is the standard logistic regression (no-covariate interaction) model linear separation boundary for a fixed threshold and the curved dashed line is the Eq. (1) (ideal) separation boundary. The sensitivity = 0.78 and false positive fraction = 0.42