search

stevenpawley / Pyspatialml

Machine learning modelling for spatial data
GNU General Public License v3.0
145 stars 29 forks source link

Compatible API question #21

Closed bluetyson closed 4 years ago

bluetyson commented 4 years ago

Hi Steven,

This looks really nice. Reading the info, it mentions something with fit and predict should work. I am interested in utilising the predict part on a raster stack with a model from elsewhere.

For example, NGBoost :- https://github.com/stanfordmlgroup/ngboost

Which is designed to be scikit-learn style.

Thanks,

Richard

bluetyson commented 4 years ago

Apparently, works very nicely!

NGBoost also has a pred_dist function to get the stdev parameter for example, which also made a fairly straightforward addition.

e.g. result_cla = estimator.pred_dist(flat_pixels) result_cla = result_cla.params['scale']

RichardScottOZ commented 3 years ago

see pull request for implementation