. 2016 Dec 23;17(1):18. doi: 10.3390/s17010018

Table 2.

kNN, NBC and SVM configuration parameters.

Configuration	kNN	NBC	SVM
Method	Euclidean, Mahalanobis	Gauss ⁽²⁾, KSF ⁽³⁾	Linear ⁽⁴⁾, quadratic ⁽⁵⁾
Data normalisation	Yes ⁽¹⁾	No	Yes ⁽¹⁾
Metrics	LOOCV, ROC	LOOCV, ROC	LOOCV, ROC
Classes	2	2	2

⁽¹⁾ Normalised based on $(x_{i} - μ) / σ$ ; ⁽²⁾ Kernel Smoothing Function: G(x) = ((2 * $π$ )^−0.5) * exp(−0.5 * x²); ⁽³⁾ Kernel Smoothing Function: $f_{h} (x) = (\frac{1}{n * h}) \sum_{i = 1}^{n} K (\frac{x - x_{i}}{h})$ ; ⁽⁴⁾ Kernel Smoothing Function: G(x₁,x₂) = x₁’x₂; ⁽⁵⁾ Kernel Smoothing Function: G(x₁,x₂) = (1 + x₁’x₂)².