A clinic wants to predict how many minutes each patient will wait before being seen. Three versions of the model were trained on the same patients. Each was then checked on a validation set, a group of patients held back from training. Every number below is the model's average error, in minutes:
| Model | Training error | Validation error |
|---|---|---|
| P | 14.2 | 14.6 |
| Q | 6.1 | 6.9 |
| R | 0.4 | 12.8 |
For scale, simply predicting the clinic's average wait for every patient is off by about 15 minutes on both sets.
Which diagnosis of the three models is correct?
Select all that apply.