Weighted Bayesian Belief Network: A Computational Intelligence Approach for Predictive Modeling in Clinical Datasets

Table 3

Generation of strong rules on basis of WBC and WBL and its accuracy.

S.No	Minimum threshold	Training dataset (%)	Testing dataset (%)	No. of association rules based on weighted support and weighted confidence	No. of strong association rules based on WBC and WBL	Accuracy (%)
1	Support = 36% Confidence = 70%	100	100	22	10	97.08
2		80	20	11	7	95.7
3		70	30	11	5	97.18
4		60	40	11	7	92.5
5	Support = 40% Confidence = 80%	100	100	11	7	89.53
6		80	20	11	7	95.74
7		70	30	11	7	86
8		60	40	28	12	92.55
9	Support = 26% Confidence = 60%	100	100	23	11	89.53
10		80	20	22	11	95.74
11		70	30	23	12	97.18
12		60	40	11	9	92.55
13	Support = 10% Confidence = 50%	100	100	23	12	89.53
14		80	20	23	12	95.74
15		70	30	23	12	97
16		60	40	23	12	92.5