Research Article
Early Diagnosis of Tuberculosis Using Deep Learning Approach for IOT Based Healthcare Applications
Table 1
Performance of proposed DBN-AMBO with other classifiers.
| Datasets | Performance | RNN | CNN | GAN | DBN | DBN-AMBO |
| Shenzhen China | Accuracy | 0.895 | 0.792 | 0.972 | 0.973 | 0.992 | Precision | 0.893 | 0.927 | 0.970 | 0.981 | 0.978 | Recall | 0.935 | 0.858 | 0.932 | 0.934 | 0.954 | Specificity | 0.961 | 0.718 | 0.856 | 0.961 | 0.991 | NPV | 0.214 | 0.347 | 0.175 | 0.283 | 0.06 | FNR | 0.850 | 0.923 | 0.971 | 0.856 | 0.998 |
| Montgomery Country | Accuracy | 0.983 | 0.894 | 0.952 | 0.915 | 0.987 | Precision | 0.853 | 0.914 | 0.906 | 0.913 | 0.966 | Recall | 0.953 | 0.884 | 0.926 | 0.954 | 0.989 | Specificity | 0.862 | 0.776 | 0.871 | 0.964 | 0.994 | NPV | 0.221 | 0.343 | 0.074 | 0.327 | 0.02 | FNR | 0.852 | 0.913 | 0.965 | 0.931 | 0.972 |
|
|