Table 2.

Comparison of predictive performances.

Method	Target direction	Average over 10-fold cross validation
		Total	TP	FN	FP	TP	Hold	Accuracy (%)	Sensitivity (%)	Specificity (%)
CBA	Inc	14.9	4.4	1.1	1.4	8	–	83	82	85
LDA	Inc	14.9	2.7	1	2.8	8.4	–	75	72	75
CBA-DR	Inc	14.9	4.4	0	1.4	0.8	8.3	79	100	29
CBA	Dec	14.9	0.2	0.7	1.4	12.6	–	86	22	90
LDA	Dec	14.9	0.2	3.3	0.7	10.7	–	73	6	95
CBA-DR	Dec	14.9	0	0.7	0	12.6	1.6	95	0	100

Predictive performance of classifiers was compared among CBA, LDA, CBA-DR with 10-fold cross validation.

Target direction: a classifier was built for whether increased (Inc) or decreased (Dec) relative liver weight. Total: average number of total records in a test set of each trial in a cross validation. TP: average number of true positive records in a test set. FN: average number of false negative records in a test set. FP: average number of false positive records in a test set. TN: average number of true negative records in a test set. Hold: average number of records in a test set that did not match any rules except the default rule (only for CBA-DR).

Note that accuracy, sensitivity and specificity for the CBA-DR method were calculated excluding ‘hold’ samples. Totals are not integers here, as the number of records in the original dataset was 149 and thus cannot be divided by 10, the number of trials in the cross validation.