Random forest demo
- Source Code rf_stack.py
- S2 image and BCGW groundref are in stack.tar.gz which rf_stack.py will auto-extract (isn't it fun to read data without a driver?)
- 20% of pixels were selected for training
- seven estimators were used
- surprisingly, the qualitative features of some of the ground-ref data seem to be decently recovered..
To run:
python3 -m "pip install scikit-learn"
python3 rf_stack.py
