Feat/xgboost

[hrzn]

An XGBoost implementation, supporting both poisson and quantile likelihoods. For now it still relies on MultiOutputRegressor for simplicity (this way it reuses what we have in place for other regression models); we can try and leverage multi-output capabilities of xgboost at some point.

Solves #1134

[codecov-commenter]

Codecov Report

Base: 93.97% // Head: 94.00% // Increases project coverage by +0.02% 🎉

Coverage data is based on head (5a57c98) compared to base (e9f850d).
Patch coverage: 98.48% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1405      +/-   ##
==========================================
+ Coverage   93.97%   94.00%   +0.02%     
==========================================
  Files          82       83       +1     
  Lines        8915     8968      +53     
==========================================
+ Hits         8378     8430      +52     
- Misses        537      538       +1     

Impacted Files	Coverage Δ
darts/models/forecasting/catboost_model.py	100.00% <ø> (ø)
darts/models/forecasting/lgbm.py	100.00% <ø> (ø)
darts/models/forecasting/xgboost.py	98.46% <98.46%> (ø)
darts/utils/multioutput.py	82.14% <100.00%> (ø)
darts/timeseries.py	91.94% <0.00%> (-0.06%)	⬇️
...arts/models/forecasting/torch_forecasting_model.py	87.68% <0.00%> (-0.06%)	⬇️
darts/models/forecasting/block_rnn_model.py	98.24% <0.00%> (-0.04%)	⬇️
darts/models/forecasting/nhits.py	99.27% <0.00%> (-0.01%)	⬇️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[eliane-maalouf]

Thanks for this implementation 👍 I added some basic remarks.
(PS: I did not run tests myself)