. 2021 Aug 17;11:16682. doi: 10.1038/s41598-021-95957-w

Table 5.

The comparison results between DF-COVID-19 and other studies.

Model	Accuracy	AUC	Sensitivity	Specificity
DF-COVID-19 with⁶¹ features	99.50% [95%CI 99.09–100]	99.91% [95%CI 99.7–100]	95.2% [95%CI 90.9–98.8]	99.96% [95%CI 99.7–100]
ER-CoV⁶¹	–	86.78% [95%CI 85.65–87.9]	70.25% [95%CI 66.57– 73.12]	85.98% [95%CI 84.94–86.84]
DF-COVID-19 with⁶² features	99.48% [95%CI 98.98–100]	99.91% [95%CI 99.3–100]	95.23% [95%CI 90.1–99.0]	99.94% [95%CI 99.7–100]
ANN with SMOTE⁶²	87%	80 ± 0.09	43%	91%
DF-COVID-19 with⁶³ features	99.49% [95%CI 98.98–100]	99.95% [95%CI 99.8–100]	95.28% [95%CI 89.8–99.0]	99.95% [95%CI 99.6–100]
Bayes Net⁶³	95.159 ± 0.693	–	96.8 ± 0.007	93.6 ± 0.011
DF-COVID-19 with⁶⁴ features	99.45% [95%CI 98.87–100]	99.91% [95%CI 99.6–100]	95.06% [95%CI 89.9–98.8]	99.92% [95%CI 99.6–100]
SVM⁶⁴	–	85%	86%	85%