Construction and Optimization of Fuzzy Rule-Based Classifier with a Swarm Intelligent Algorithm

Table 4

Comparison between the initial fuzzy classifier and the optimized one for the Iris data set.

Data set name	Fuzzy classification system (FCS)	The number of feature variables	The number of fuzzy sets	The number of fuzzy rules	The number of correctly classified samples	Classification accuracy (%)
Iris	Initial FCS	2	8	16	143	95.33
	Optimized FCS	2	8	9	143	95.33
	C4.5	2	8	9	138	92.00
Wine	Initial FCS	5	10	32	161	90.45
	Optimized FCS	5	10	10	161	90.45
	C4.5	5	10	10	159	89.32
Glass	Initial FCS	5	10	32	180	84.11
	Optimized FCS	5	10	17	180	84.11
	C4.5	5	10	17	171	81.42
Seeds	Initial FCS	5	15	243	174	82.86
	Optimized FCS	5	15	12	174	82.86
	C4.5	5	15	12	167	79.52
Wdbc	Initial FCS	10	20	1000	451	79.31
	Optimized FCS	10	20	97	451	79.31
	C4.5	10	20	97	436	76.62
Satimage	Initial FCS	10	20	1000	5246	81.53
	Optimized FCS	10	20	153	5246	81.53
	C4.5	10	20	153	4654	72.32