Table 5

Percentiles relating to bias of repeated examinations of a test set

Percentile


10th

25th

50th

75th

90th


DS1, m/z > 0

Bootstrap smallest error:

0.0

0.0

0.0

0.0

0.0

Original cohort error:

.01

.01

.02

.03

.05

Estimated bias:

.003

.012

.020,

.032

.051

DS1, m/z > 1500

Bootstrap smallest error:

.02

.02

.03

.05

.06

Original cohort error:

.07

.07

.08

.11

.13

Estimated bias:

.02

.04

.05,

.07

.10

DS2, 700 <m/z < 12000

Bootstrap smallest error:

.01

.02

.03

.04

.05

Original cohort error:

.04

.05

.06

.08

.10

Estimated bias:

.01

.03

.04,

.06

.07


Jeffries BMC Bioinformatics 2004 5:180   doi:10.1186/1471-2105-5-180

Open Data