Experimental
Random Forest regression for continuous targets (e.g., biomass, moisture, temperature) from raster predictors.
remote_sensing regression random_forest legacy-port
| Name | Description | Required | Default |
|---|---|---|---|
inputs | Array of single-band input rasters. | Required | ['band1.tif', 'band2.tif', 'band3.tif'] |
training_data | Point vector training data path. | Required | training_points.shp |
field | Numeric target field in training_data attributes. | Required | value |
scaling | Feature scaling mode: none (default), normalize, standardize. | Optional | none |
n_trees | Number of trees in the forest (default 200). | Optional | 200 |
min_samples_leaf | Minimum number of samples required at a leaf node (default 1). | Optional | 1 |
min_samples_split | Minimum number of samples required to split an internal node (default 2). | Optional | 2 |
output | Optional output raster path. | Optional | — |
Run random forest regression on multiband predictors.
wbe.random_forest_regression(field='target', inputs=['band1.tif', 'band2.tif', 'band3.tif'], n_trees=300, output='rf_regression.tif', scaling='standardize', training_data='training_points.shp')