Code and data for the paper "Instance-based Transfer Learning for Soil Organic Carbon Estimation"
Here you can find a zip file with code and data used to perform the experiments from the paper.
When you unizp the file, there will be three folders and two Python files. Folders contain source domain grassland and cropland data and target domain data. Each of 12 countries is represented with its training-testing split.
Python scripts perform the experiments from the paper (tl_github.py) and calculate Bhattasharyya distances between the source and target distributions (dist_github.py). You need to install pytorch, numpy and sklearn libraries to run the code!
For any question please send email to milos@grf.bg.ac.rs :)