Table 7 Comparative accuracies with respect to the seven-classication application and the eight-classication application of whether to use class weights
From: Intelligent diagnosis with Chinese electronic medical records based on convolutional neural networks
Class ∖metrics | Name of class | Sample size | Seven-classication | Eight-classication | ||
|---|---|---|---|---|---|---|
Without class weight | With class weight | Without class weight | With class weight | |||
Class1 | Allergic rhinitis | 1079 | 71.09 | 80.1 | 59.68 | 77.85 |
Class2 | Respiratory disease | 11980 | 90.37 | 87.92 | 85.28 | 86.3 |
Class3 | Cough variant asthma | 1418 | 70.31 | 80.74 | 67.12 | 81.45 |
Class4 | Acute bronchitis | 11990 | 77.5 | 80.00 | 65.56 | 81.78 |
Class5 | Bronchialasthma, no critical | 1550 | 79.23 | 83.56 | 78.82 | 80.77 |
Class6 | Bronchitis | 17726 | 82.79 | 73.42 | 66.94 | 51.42 |
Class7 | Diarrhea | 3405 | 97.91 | 98.7 | 94.9 | 97.06 |
Class8 | Acute upper respiratory tract infection | 43596 | NA | NA | 92.94 | 84.11 |