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

Table 5

The comparison among the fuzzy classifiers generated by different methods for the four data sets.

Modeling method	Data set	The number of fuzzy rules	The number of the fuzzy rules after optimized	The number of correctly classified samples	Sample size	The classification accuracy (%)
FCS directly established from the samples	Iris		26	143	150	95.33
	Wine		70	170	178	95.51
	Glass		118	188	214	87.85
	Seeds		34	191	210	90.95
	Wdbc		102	441	569	77.51
	Satimage		157	5117	6435	78.31
FCS established and optimized with GA	Iris		12	142	150	94.67
	Wine		12	160	178	89.89
	Glass		17	170	214	79.43
	Seeds		15	183	210	87.14
	Wdbc		108	440	569	77.43
	Satimage		133	5061	6435	78.65
FCS established and optimized with BPSO	Iris		12	142	150	94.67
	Wine		12	160	178	89.89
	Glass		17	173	214	80.84
	Seeds		12	185	210	88.10
	Wdbc		104	418	569	73.53
	Satimage		160	4932	6435	76.65
FCS established and optimized with QiQPSO	Iris		9	143	150	95.33
	Wine		10	161	178	90.45
	Glass		17	180	214	84.11
	Seeds		12	190	210	90.47
	Wdbc		200	449	569	78.91
	Satimage		153	5189	6435	80.65