Use of Attribute Driven Incremental Discretization and Logic Learning Machine to build a prognostic classifier for neuroblastoma patients

Table 2 Classification rules.

Rule ID^a		Cond 1	Cond 2		Predicted Outcome	Covering^b (%)	Error^c (%)	Fisher pvalue^d
1	IF(	217356_s_at ≤ 721	226452_at < 326	)THEN	Good	80	3.5	<0.001
2	IF (	206686_at ≤ 26	226452_at ≤326	)THEN	Good	70	14	<0.001
3	IF (	200738_s_at ≤ 1846	230630_at> 23	)THEN	Good	62	10	<0.001
4	IF(	209446_s_at ≤ 57	223172_s_at < 73	)THEN	Good	60	10	<0.001
5	IF(	202022_at > 131	223193_x_at < 324	)THEN	Good	60	14	<0.001
6	IF(	224314_s_at ≤ 29	236180_at <13	)THEN	Good	48	7.1	<0.001
7	IF(	217356_s_at > 721		)THEN	Poor	92	17	<0.001
8	IF(	223172_s_at >73	226452_at> 326	)THEN	Poor	60	8.6	<0.001
9	IF(	206686_at > 26	223172_s_at>73	)THEN	Poor	57	7.4	<0.001

^a Cond 1 and Cond 2 indicate the conditions into the premises of the rules.
^b The covering accounts for the fraction of patients that verify the rule and belong to the target outcome.
^c The error accounts for the fraction of patients that satisfy the rule and do not belong to the target outcome.
^d Fisher p-value quantifies the statistical significance of the rule on the basis of the number of patients correctly and incorrectly classified by a rule and the number of patients of the dataset belonging to each specific outcome.

ISSN: 1471-2105