Statistical analysis of a Bayesian classifier based on the expression of miRNAs

BMC Bioinformatics

Table 3 Statistics of Δ x ₂₀₅, Δ x ₂₁, x _U6, y _DV (see Eq. (2)), y _opt (see Eq. (10))

Measure	Class	Set size	\(\overline {X}\)	S	TN	t-test
Δ x ₂₀₅	target	18	1.7(5)	2.2(4)	0.71	2·10⁻⁸
	versus	15	–4.6(6)	2.4(5)	0.26
Δ x ₂₁	target	19	–6.5(3)	1.4(2)	0.53	0.90
	versus	16	–6.6(6)	2.4(4)	0.94
x _U6	target	19	25.0(3)	1.4(2)	0.40	0.14
	versus	16	25.8(5)	1.9(3)	0.25
y _DV	target	18	4.9(5)	2.0(3)	0.96	5.5·10⁻¹¹
	versus	15	–1.3(4)	1.7(3)	0.98
y _opt	target	18	6.9(5)	1.9(3)	0.78	2.6·10⁻¹¹
	versus	15	0.7(4)	1.6(3)	0.24

The digits in parentheses correspond to the uncertainty on the respective least significant digits. The column marked with TN (test of normality) contains the p-values yielded by the Shapiro-Wilk test to check whether the data contained in a histogram are consistent with a normally distributed parent population. Finally, the rightmost column reports the p-value of Student’s t-test to check the null hypothesis that the target and versus sets have the same population mean; to this purpose, the variance is estimated separately for each group and the Welch modification to the number of degrees of freedom is used