Table 1.
Variables considered for inclusion in the Bayesian network model.
Fig 1.
Final Bayesian network model for dengue diagnosis.
Table 2.
Demographic characteristics of acute undifferentiated febrile patients recruited in the study.
Table 3.
Final diagnosis of dengue and non-dengue patients.
Table 4.
Performance of different dengue diagnosis models, determined by Area Under the Curve (AUC) of ROC analysis.
Table 5.
Mean and standard deviation of AUC using 10-fold cross-validation.
Table 6.
Overall accuracy of physician’s dengue diagnosis without NS1 rapid test result.
Table 7.
Overall accuracy of physician’s dengue diagnosis with NS1 rapid test result.
Table 8.
Overall performance on dengue diagnosis of BN model and physician.