Configure Auto Learn Threshold
To configure Auto Learn Threshold per test, do the following:
-
Turn on Auto Learn Threshold for the test.
-
Configure the Auto Learn Tolerance percentile. The percentile is based on the probability that a measurement value can be abnormal. The percentile can be estimated from observing the data, or from domain expertise, or a combination of the two. For example, if you estimate that one in every thousand measurements is an abnormal, set the percentile to 99.9 by using the formula (1–1/1000) * 100. See the following table for examples:
| Outlier Frequency | Percentile | Formula |
|---|---|---|
| 1 in 100,000 | 99.999 | (1–1/100000) * 100 |
| 1 in 1,000,000 | 99.9999 | (1–1/1000000) * 100 |
| 5 in 100,000 | 99.995 | (1–5/100000) * 100 |
| 3 in 1,000,000 | 99.9997 | (1–3/1000000) * 100 |
| P in Q | See the Formula column | (1–P/Q) * 100 |
The default value is Medium sensitivity Percentile 99.
