Table 6.

Comparison in terms of accuracies (%) and number of selected features obtained by the proposed method with some state-of-the-art methods for the arrhythmia dataset (highest accuracy and lowest #features are highlighted in bold).

Dataset	Method	No. of Features Selected	Classification Accuracy (%)
Arrhythmia	Xu et al. [11]	236	82.96
	Singh et al. [12]	30	85.58
	Sahebi et al. [13]	135	99.02
	Cui et al. [14]	169	74.77
	Kadam et al. [15]	92	88.72
	Wang et al. [16]	89	98.68
	Proposed	3	94.50