Table 1.

Error rates of algorithms in four classification problems




Decision tree

t-rank

SPLPa

SPLRa

Gaussian Kernel SVM

Neural net
Test errorb
Azoles 7.2 10.3 5.8 4.1 6.8 4.4
Fibrates 14.6 4.0 1.6 2.8 1.3 1.1
Statins 23.7 9.3 6.4 5.6 6.5 5.4
Toxicants 14.1 18.0 10.7 11.9 5.6 5.7
Average 14.9 10.4 6.1 6.1 5.0 4.2
Training error
Azoles 0.2 6.0 0.0 0.0 0.0 0.0
Fibrates 0.9 5.0 0.5 0.0 0.0 0.0
Statins 0.7 9.0 3.0 0.0 0.0 0.0
Toxicants 0.7 15.1 0.0 0.0 0.0 0.0
Average
0.6
8.8
0.9
0.0
0.0
0.0

a SPLP and SPLR were run with ρ values of 0.08 and 0.0025, respectively.

b Test error is an average rate for each classification over 20 random partitions of the data.