Data and Feature Reduction in Fuzzy Modeling through Particle Swarm Optimization

Table 1

A summary of selected studies in data and feature reduction in fuzzy modeling.

Reference	Feature reduction technique	Dataset, fuzzy model and data	Original data used in modeling		Number of selected features	Number of resulting rules
			Number of instances	Number of features
Gaweda et al. [15]	The use of sensitivity analysis; Determination of essential features	Box-Jenkins gas furnace	296	10	3	2
Hadjili and Wertz [16]	Deviation criterion (DC): to measure the change in fuzzy partition. Removal of features that do not significantly change the fuzzy partition	Nonlinear systems in noisy environment	250	3	1	4
		Nonlinear dynamical system excited by a sinusoidal signal	800	10	6	8
		Run-out cooling table in a hot strip mill	1000	17	5	12
Zarandi et al. [18]	Heuristic method to select features	Nonlinear System used in [3]	50	4	2	4
		Supplier chance management dataset	300	9	5	5
Du and Zhang; [19]	Evolutionary optimization	Box-Jenkins gas furnace	296	10	3	4
		MR damper identification	5000	11	6	10
Ghazavi and Liao [20]	(1) Mutual; correlation methods,; (2) gene selection criteria; (3) the relief algorithm	Wisconsin breast cancer; PIMA Indian diabetes; Welding flaw identification	569; 768; 399	30; 8; 25	3; 3; 3	250 (3); 125 (3); —
Zhang et al. [21]	Iterative search margin based algorithm (Simba)	Wisconsin breast cancer	699	9	5	3
		Wine	178	13	4	5
		Iris	150	4	3	3
		Ionosphere	351	34	10	4