Skip to main content
. 2012 Feb 22;7(2):e32383. doi: 10.1371/journal.pone.0032383

Figure 1. Biomarker selection and training/testing analysis.

Figure 1

Before biomarker selection, our total sample pool was split into two randomized groups: the Training Set and Testing Set. Prospector and PAM statistical analyses were performed on the Training Set to identify the top 10 most significant autoantibody classifiers of PD and control. We then verified the diagnostic accuracy of these selected biomarkers by using Random Forest to predict sample classification in the Training Set, Testing Set, and then both sets combined.