|
GDFS
|
CFS
|
rpart
|
---|
|
Gleason 5
|
Gleason 7
|
Gleason 5
|
Gleason 7
|
Gleason 5
|
Gleason 7
|
---|
True groups
|
Gleason 5
|
11
|
6
|
11
|
6
|
11
|
6
|
|
Gleason 7
|
4
|
13
|
9
|
8
|
8
|
9
|
- Statistical classifications of the prostate cancer dataset (Gleason 5 vs. Gleason 7 cases) from the proposed GDFS method (GDFS), correlation-based feature selection (CFS), and recursive partitioning (rpart). In each case, posterior group probabilities were calculated for each observation j using features selected, and model parameters estimated, with observation j omitted (leave-one-out cross-validation). Observations were then classified using a MAP classification rule. This table shows the cross-tabulations of true group membership with the assigned classifications from the GDFS, CFS and rpart methods.