Skip to first unread message
Andrew Orry
May 25, 2017, 9:08:29 PM5/25/17
to MolSoft ICM Knowledge Base
Q.
How to calculate a ROC curve for Virtual Ligand Screening Data?\
A.
The macro calc3Rocs can be used, it takes two array: R_scores is
the rarray of score (the more negative value is the better score);
I_labels is the iarray indicating which rows are positive (1 is
positive, 0 is negative). Let's say you have a table t that has
the column Scores and category:
calc3Rocs t.Scores t.category yes "blue"
will return the ROCs table with 3 curves, the last of which is the ROC curve. R_out[1] will return the AUC of the ROC curve.
calc3Rocs t.Scores t.category yes "blue"
will return the ROCs table with 3 curves, the last of which is the ROC curve. R_out[1] will return the AUC of the ROC curve.
Reply all
Reply to author
Forward
0 new messages