Research Article
A Mobile-Based Question-Answering and Early Warning System for Assisting Diabetes Management
Table 3
The comparison of our answer selection method with baseline methods using different k settings.
| Setting | Methods | Accuracy | Precision | Recall | F1 |
| k=5 | KNN | 0.8052 | 0.6820 | 0.5679 | 0.5746 | GaussianNB | 0.7383 | 0.6980 | 0.8026 | 0.6956 | RF | 0.8301 | 0.7339 | 0.7097 | 0.7203 | SVM | 0.8373 | 0.7513 | 0.8029 | 0.7706 | PPN | 0.8842 | 0.8129 | 0.8540 | 0.8305 | Our method | 0.9222 | 0.8859 | 0.8657 | 0.8753 |
| k=10 | KNN | 0.9032 | 0.8076 | 0.5258 | 0.5247 | RF | 0.9106 | 0.7523 | 0.7277 | 0.7391 | GaussianNB | 0.8956 | 0.7197 | 0.7737 | 0.7422 | SVM | 0.9154 | 0.7642 | 0.8399 | 0.7951 | PPN | 0.9271 | 0.7897 | 0.8633 | 0.8206 | Our method | 0.9651 | 0.9415 | 0.8559 | 0.8929 |
|
|