Probabilistic factor models for spatially correlated data. Includes nonnegative and real-valued versions.
Many people find it convenient to install everything in a conda environment. We recommend installing the miniforge version of conda. You can then create an environment and activate it with the following commands:
conda create -n nsf python=3.11
conda activate nsf
The package can then be installed into the environment from github:
pip install git+https://github.com/willtownes/spatial-factorization-py.git#egg=spatial-factorization
import spatial_factorization
help(spatial_factorization.ModelTrainer)
help(spatial_factorization.SpatialFactorization)
First make sure that the right tools are installed:
pip install hatch
pip install keyrings.alt
Then build with hatch:
hatch build