. 2020 Oct 26;10:18291. doi: 10.1038/s41598-020-75500-z

Table 2.

Leave-one-out cross-validation assessment and independent cohort validation in the four patient groups.

Prediction Leave-one-out cross-validation	Viral encephalitis	Bacterial meningitis	Acquired Immune	Control group
Viral encephalitis (n = 107)	96 89.7%	2 1.87%	7 6.54%	2 1.87%
Bacterial meningitis (n = 15)	4 26.7%	9 60%	1 6.67%	1 6.67%
Acquired Immune (n = 48)	15 31.2%	0 0%	13 27.1%	20 41.7%
Control group (n = 107)	1 0.93%	0 0%	7 6.54%	99 92.5%
Independent cohort validation
Viral encephalitis (n = 9)	7 77.8%	1 11.1%	1 11.1%	0 0%
Bacterial meningitis (n = 4)	2 50%	2 50%	0 0%	0 0%
Acquired immune (n = 8)	1 12.5%	0 0%	6 75%	1 12.5%
Control group (n = 65)	1 1.54%	0 0%	4 6.15%	60 92.3%
Genetic diseases (n = 9)	5 55.6%	3 33.3%	0 0%	1 11.1%

Prediction Leave-one-out cross-validation

Viral encephalitis

Bacterial meningitis

Acquired Immune

Control group

Viral encephalitis

(n = 107)

89.7%

1.87%

6.54%

1.87%

Bacterial meningitis

(n = 15)

26.7%

60%

6.67%

Acquired Immune

(n = 48)

31.2%

27.1%

41.7%

Control group

(n = 107)

0.93%

6.54%

92.5%

Independent cohort validation

Viral encephalitis

(n = 9)

77.8%

11.1%

Bacterial meningitis

(n = 4)

50%

50%

Acquired immune

(n = 8)

12.5%

75%

12.5%

Control group

(n = 65)

1.54%

6.15%

92.3%

Genetic diseases

(n = 9)

55.6%

33.3%

11.1%

Data are expressed as number of cases and the percentage of right classifications obtained. The highest percentage of correct classification was reached for the viral encephalitis and control groups in both assessments (percentages highlighted in bold), followed by the acquired immune and bacterial meningitis groups. Patients from the group with genetic immune disease (Aicardi-Goutières), we classified as viral or bacterial meningitis and only one as a control.