Target Class Options

Reference > Target Class Options

ecoClassify trains a binary (presence/absence) classifier. The Target Class Options Datasheet lets you train from a multi-class user-classified raster by choosing one class as the target. When it is not used, the lowest value in each user-classified raster is treated as absence (0) and the highest as presence (1).

Field Description
Use target class When checked, converts the user-classified rasters to presence/absence using the fields below.
Target class value The class code to predict. Pixels with this value become presence (1).
Background values (comma-separated) Optional, for example 2,5,7. When given, only these class codes become absence (0); all other codes are excluded from training. If a value is listed here and is also the target, it is treated as background. When left blank, every class other than the target (and the ignore values) becomes absence.
Ignore values (comma-separated) Optional. Class codes excluded from training, such as clouds or unclassified pixels. Used only when Background values is blank.
Target class label Optional. A readable name for the target class, such as Snow, used in the evaluation outputs.