Table 3 Number of tokens/unique tokens per data set and split
From: Classifying patient and professional voice in social media health posts
| Both data sources | |||
|---|---|---|---|
| Cardiovascular | |||
| Train | 831,169/26,037 | 119,087/16,118 | 950,256/34,998 |
| Test | 211,486/13,302 | 30,257/6729 | 241,743/17,094 |
| Skin | |||
| Train | 1,159,225/29,176 | 98,410/13,639 | 1,257,635/35,854 |
| Test | 290,227/14,201 | 24,337/5483 | 314,564/16,731 |
| Both domains | |||
| Train | 1,990,394/43,390 | 217,497/25,441 | 2,207,891/57,118 |
| Test | 501,713/21,779 | 54,594/10,201 | 556,307/26,444 |