Table 3 Prediction on all the hospital visits on the test dataset with a confidence of 93%, 95%, and 99% compared to the clinical diagnosis

From: Conformal prediction enables disease course prediction and allows individualized diagnostic uncertainty in multiple sclerosis

Predictions

93% confidence

95% confidence

99% confidence

RRMS (nv = 18,525)

SPMS (nv = 9798)

RRMS (nv = 18,525)

SPMS (nv = 9798)

RRMS (nv = 18,525)

SPMS (nv = 9798)

RRMS

19,616 (91.7%)

401 (5.8%)

19,351 (90.5%)

269 (3.9%)

18,371 (85.9%)

67 (1%)

SPMS

1672 (7.8%)

6484 (93.5%)

1168 (5.5%)

6116 (88.2%)

315 (1.5%)

4569 (65.9%)

Empty-{}

0 (0%)

0 (0%)

0 (0%)

0 (0%)

0 (0%)

0 (0%)

Multiple-label (RRMS | SPMS)

100 (0.5%)

50 (0.7%)

869 (4.1%)

550 (7.9%)

2702 (12.6%)

2299 (33.2%)

  1. The predictions of RRMS and SPMS indicate single-label prediction, empty represents no valid prediction, and multiple-label represents both RRMS and SPMS prediction for the hospital visit. nv = number of hospital visits. *Percentages would not add up due to rounding off.
Back to article page