TreeSurgeon: Visualisation of Random Forest Regressors

TreeSurgeon contains routines to visualise Random Forest Regressor models. The module takes models output files made by sklearn's RandomForestRegressor implementation of the random forest regressor algorithm. The raw output files from sklearn models (*pkl) first needs to be converted to the input .csv files required by TreeSurgeon using the extract_models4TreeSurgeon.py script in the sparse2spatial module.

Quick Start

Running

• Process the saved Random Forest Regressor models *.pkl files into the .csv that TreeSurgeon* expects using the script in sparse2spatial module. You will need to update some lines in the script as described there.

python extract_models4TreeSurgeon.py

• Place files in the csv folder.

for composite files:

python start.py $NCPUS

or for single dot files

python start.py $NCPUS 1

• This then runs in the background (no screen). To change edit show option in main.js

Set colours

The colours are set in the colours.json file.

Output

This is in the pdfs folder.

Install

conda install nodejs
npm install
sudo npm install -g --save electron --unsafe-perm=true --allow-root

• for merge imagemagick and ghostscript need to be installed

Montage setup

python montage.py

Example Output for Composite Graph

Usage

This package was initially written for use with the sparse2spatial package for work to predict sea-surface concentrations [Sherwen et al. 2019]. However it can be used for any Radom Forest Regressor models made by sklearn and post-processed to TreeSurgeon input by sparse2spatial

Reference

Sherwen, T., Chance, R. J., Tinel, L., Ellis, D., Evans, M. J., and Carpenter, L. J.: A machine learning based global sea-surface iodide distribution, Earth Syst. Sci. Data Discuss., https://doi.org/10.5194/essd-2019-40, in review, 2019.

License

Copyright (c) 2019 Daniel Ellis and Tomas Sherwen

This work is licensed under a permissive MIT License.