Table 3.

Settings of Algorithm Options

Nr	RIPPER	RIDOR	PART
1	-F 3 -N 2.0 -O 10	-F 3 -S 1 -N 2.0 -A	-R -B -M 2 -N 3
2	-F 3 -N 5.0 -O 10	-F 3 -S 1 -N 5.0 -A	-R -B -M 5 -N 3
3	-F 3 -N 10.0 -O 10	-F 3 -S 1 -N 10.0 -A	-R -B -M 10 -N 3
4	-F 10 -N 2.0 -O 10	-F 10 -S 1 -N 2.0 -A	-R -B -M 2 -N 10
5	-F 10 -N 5.0 -O 10	-F 10 -S 1 -N 5.0 -A	-R -B -M 5 -N 10
6	-F 10 -N 10.0 -O 10	-F 10 -S 1 -N 10.0 -A	-R -B -M 10 -N 10
7	-F 100 -N 2.0 -O 10	-F 20 -S 1 -N 2.0 -A	-R -B -M 2 -N 100
8	-F 100 -N 5.0 -O 10	-F 20 -S 1 -N 5.0 -A	-R -B -M 5 -N 100
9	-F 100 -N 10.0 -O 10	-F 20 -S 1 -N 10.0 -A	-R -B -M 10 -N 100
10	-F 3 -N 2.0 -O 100		-R -M 2 -N 3
11	-F 3 -N 5.0 -O 100		-R -M 5 -N 3
12	-F 3 -N 10.0 -O 100		-R -M 10 -N 3
13	-F 10 -N 2.0 -O 100		-R -M 2 -N 10
14	-F 10 -N 5.0 -O 100		-R -M 5 -N 10
15	-F 10 -N 10.0 -O 100		-R -M 10 -N 10
16	-F 100 -N 2.0 -O 100		-R -M 2 -N 100
17	-F 100 -N 5.0 -O 100		-R -M 5 -N 100
18	-F 100 -N 10.0 -O 100		-R -M 10 -N 100
19			-B -M 2 -C 0.25
20			-B -M 2 -C 0.1
21			-B -M 5 -C 0.25
22			-B -M 5 -C 0.1
23			-B -M 10 -C 0.25
24			-B -M 10 -C 0.1
25			-M 2 -C 0.25
26			-M 2 -C 0.1
27			-M 5 -C 0.25
28			-M 5 -C 0.1
29			-M 10 -C 0.25
30			-M 10 -C 0.1

RIPPER: F: number of folds for reduced error pruning; N: minimal weights of instances within a split; O: number of optimization runs

RIDOR: F: number of folds for reduced error pruning; S: number of shuffles for randomization; A: Flag set to use the error rate of all the data to select the default class in each step. N: minimal weight of instances within a split.

PART: C: confidence threshold for pruning; M: minimum number of instances per leaf; R: use reduced error pruning; N: number of folds for reduced error pruning; B: Use binary splits for nominal attributes