Table 5 Model performance indices of the LR, DT, BF algorithms for Model I, II, and III in training data
From: Association between biochemical and hematologic factors with COVID-19 using data mining methods
Model I | ||||||
|---|---|---|---|---|---|---|
(a) LR | (b) DT | |||||
Actual | Predicted Count | Actual | Predicted Count | |||
COVID Positive | No | Yes | COVID Positive | No | Yes | |
No | 3328 | 675 | No | 5149 | 758 | |
Yes | 1075 | 1959 | Yes | 2061 | 2568 | |
Sensitivity = 83.14% | AUC = 80.74% | Sensitivity = 87.17% | AUC = 80.23% | |||
Precision = 75.58% | Accuracy = 75.13% | Precision = 71.41% | Accuracy = 73.24% | |||
(c) BF | Â | |||||
Actual | Predicted Count | |||||
COVID Positive | No | Yes | ||||
No | 5718 | 189 | ||||
Yes | 819 | 3810 | ||||
Sensitivity = 96.80 % | AUC = 98.06 % | Â | Â | |||
Precision = 87.47 % | Accuracy = 90.43 % | Â | Â | |||
Model II | ||||||
(d) LR | (e) DT | |||||
Actual | Predicted Count | Actual | Predicted Count | |||
COVID Positive | No | Yes | COVID Positive | No | Yes | |
No | 4074 | 1175 | No | 4506 | 1401 | |
Yes | 1764 | 1175 | Yes | 1401 | 2925 | |
Sensitivity = 77.61 % | AUC = 77.37 % | Sensitivity = 76.28 % | Sensitivity = 76.28 % | |||
Precision = 69.78 % | Accuracy = 68.28 % | Sensitivity = 76.28 % | Sensitivity = 76.28 % | |||
(f) BF | Â | |||||
Actual | Predicted Count | |||||
COVID Positive | No | Yes | ||||
No | 5488 | 419 | ||||
Yes | 1262 | 3367 | ||||
Sensitivity = 92.91 % | Precision = 81.30 % | Â | Â | |||
Precision = 81.30 % | Accuracy = 84.05 % | Â | Â | |||
Model III | ||||||
(g) LR | (h)DT | |||||
Actual | Predicted Count | Actual | Predicted Count | |||
COVID Positive | No | Yes | Predicted Count | No | No | |
No | 3890 | 871 | No | 5282 | 625 | |
Yes | 1273 | 2427 | Yes | 2176 | 2453 | |
Sensitivity = 66.08% | AUC = 80.37 % | Sensitivity = 66.00% | Precision = 72.88% | |||
Precision = 74.93% | AUC = 80.37 % | Precision = 72.88% | Precision = 72.88% | |||
(i) BF | Â | |||||
Actual | Predicted Count | |||||
COVID Positive | No | Yes | ||||
No | 5808 | 99 | ||||
Yes | 647 | 3982 | ||||
Sensitivity = 66.08% | AUC = 99.00 % | Â | Â | |||
Precision = 74.93% | Accuracy = 69.63% | Â | Â | |||