Table 3.

Implementation parameters for feature selection methods

Feature selection method	Implementation
Fast Correlation Based Filter (FCBF)	• Decreasing order of relevance
	• ${10}^{-4}$ SU threshold
	• FEAST Toolbox for Matlab® (Brown, Pocock, Zhao, & Luján, 2012)
Information Gain (IG)	• Decreasing order of relevance
	• Forward selection – first order utility (Brown, 2009)
Relief	• Decreasing feature weight
	• 10 nearest neighbours
	• Matlab® function
Chi-square	• Decreasing order of ${\chi }^2$ (chi-square) value
Symmetrical uncertainty (SU)	• Decreasing order of SU value
Correlation-based Feature Selection (CFS)	• Decreasing order of heuristic merit $M_s$ for each feature subset $S$ with $k$ features ($r_{cf}$ and $r_{ff}$ represent the average feature-class and feature-feature correlations, respectively) (Hall & Holmes, 2003): $M_S=\frac{k\cdot \overline{r_{cf}}}{\sqrt{(k+k(k-1)\overline{r_{ff}}}}$
	• Correlation based on symmetrical uncertainty (SU)
Minimal Redundancy Maximal Relevance (mRMR)	• Decreasing order of relevance
	• FEAST Toolbox for Matlab® (Brown et al., 2012)