Installation
You can install the package from PyPI with the command pip install medil
.
The default only requires SciPy and allows for learning causal factor models from data, random generation of causal factor models, and sampling data from a given (or learned) causal factor model, all in the linear Gaussian setting.
There are additional optional requirements for extra features:
Key |
Description |
|
Installs all optional dependencies, for full functionality; equivalent to using all of the other keys below. |
|
Uses deep generative models for the causal mechanisms (instead of restricting to the linear Gaussian setting), specifically using PyTorch to implement a variational autoencoder. |
|
Uses NetworkX, Matplotlib and seaborn for vizualize the causal factor graph and various plots. |
In order to use these features, simply append the corresponding comma-separated keys after the package name in brackets (with no spaces), e.g., pip install medil[vis]
.