464
O. Ogorodnyk et al.
Fig. 32.2 Prediction quality of tested algorithms
Table 32.2 Performance of the tested algorithms
Perf. meas
REPTree
Random Forest
kNN
MLP
Thickness 1
Corr. coef
0.9566
0.9547
0.9226
0.9203
MAR
0.0172
0.0171
0.0227
0.0243
RMSE
0.0355
0.037
0.047
0.0528
Width 1
Corr. coef
0.8357
0.9271
0.8863
0.9014
MAR
0.0722
0.0499
0.0788
0.0644
RMSE
0.1172
0.0788
0.0976
0.1002
Thickness 2
Corr. coef
0.9811
0.9913
0.9893
0.977
MAR
0.0144
0.0118
0.0108
0.0198
RMSE
0.0261
0.0189
0.0194
0.0388
Width 2
Corr. coef
0.9017
0.9396
0.9159
0.9245
MAR
0.0697
0.0516
0.0621
0.0785
RMSE
0.1288
0.1012
0.1213
0.118