Table 3:
Regression coefficients demonstrating the change in top-k accuracy for each additional disease added to the disease space. We find decreasing accuracy, represented by a negative βD , across all models
| βD | Std. Err. | t-value | p-value | |
|---|---|---|---|---|
| Top-1 | -0.156 | 0.010 | -15.52 | ≤ 2e-16 |
| Top-3 | -0.104 | 0.0078 | -13.47 | ≤ 2e-16 |
| Top-5 | -0.102 | 0.0071 | -14.41 | ≤ 2e-16 |
| Top-10 | -0.092 | 0.0074 | -12.45 | ≤ 2e-16 |
| Top-20 | -0.070 | 0.0056 | -12.40 | ≤2e-16 |