3.4.4.5. Random Forest

_images/random_forest_tab.png

cloud_masking_tool Random forest

This tool allows for the Random Forest classification, based on a Band set and Training input.

ESA SNAP is required. The ESA SNAP GPT executable must be defined in External programs settings.

3.4.4.5.1. Random Forest classification

  • Select input band set input_number: select the input Band set to be classified;
  • Use checkbox MC ID checkbox C ID registry_save: if MC ID is checked, the classification is performed using the Macroclass ID (code MC ID of the signature); if C ID is checked, the classification is performed using the Class ID (code C ID of the signature);
  • Number of training samples input_number: set the number of training data (pixels) randomly used to traing the model; it should be set lower than total training input pixels;
  • Number of trees input_number: set the number of decision trees; a higher number allows for more accurate models, but it also increases the calculation time;
  • checkbox Evaluate classifier: if checked, the classifier is evaluated;
  • checkbox Evaluate feature power set Min input_number Max input_number: if checked, evaluate the power set of input features (e.g. Gini decrease), according to the contribution thereof to the model; Min and Max are used as thresholds for power sets; it can increase the calculation time;
  • checkbox Save classifier: if checked, save the classifier for later use;
  • Load classifier input_text open_file: open a previously saved classifier; if loaded, the input Band set is directly classified using this classifier;
  • reset: clear the classifier;
  • BATCH batch_tool: add this function to the Batch;
  • RUN run: select an output directory and start the classification process;