Merge branch 'master' of github.com:M66B/FairEmail

pull/194/head
M66B 4 years ago
commit 2bbdfc3e0d

@ -3468,15 +3468,15 @@ The message classifier calculates the probability a message belongs in a folder
There are two options in the miscellaneous settings which control if a message will be automatically moved into a folder, There are two options in the miscellaneous settings which control if a message will be automatically moved into a folder,
provided that auto classification is enabled for the folder: provided that auto classification is enabled for the folder:
* *Minimum class probability*: a message will only be moved when the confidence it belongs in a folder is greater than this value (default 20 %) * *Minimum class probability*: a message will only be moved when the confidence it belongs in a folder is greater than this value (default 15 %)
* *Minimum class difference*: a message will only be moved when the difference in confidence between one class and the next most likely class is greater than this value (default 50 %) * *Minimum class difference*: a message will only be moved when the difference in confidence between one class and the next most likely class is greater than this value (default 50 %)
Both conditions must be satisfied before a message will be moved. Both conditions must be satisfied before a message will be moved.
Considering the defaults option values: Considering the default option values:
* Apples 40 % and bananas 30 % would be disregarded because the difference of 25 % is below the minimum of 50 % * Apples 40 % and bananas 30 % would be disregarded because the difference of 25 % is below the minimum of 50 %
* Apples 15 % and bananas 5 % would be disregarded because the probability for apples is below the minimum of 20 % * Apples 10 % and bananas 5 % would be disregarded because the probability for apples is below the minimum of 15 %
* Apples 50 % and bananas 20 % would result in selecting apples * Apples 50 % and bananas 20 % would result in selecting apples
Classification is optimized to use as little resources as possible, but will inevitably use some extra battery power. Classification is optimized to use as little resources as possible, but will inevitably use some extra battery power.

Loading…
Cancel
Save